(Advances in Intelligent Systems and Computing 623) Marek Gzik, Ewaryst Tkacz, Zbigniew Paszenda, Ewa Piętka (Eds.) - Innovations in Biomedical Engineering-Springer International Publishing (2018)

Advances in Intelligent Systems and Computing 623
Marek Gzik
Ewaryst Tkacz
Zbigniew Paszenda
Ewa Piętka Editors
Innovations
in Biomedical
Engineering
Advances in Intelligent Systems and Computing
Volume 623
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
About this Series
The series “Advances in Intelligent Systems and Computing” contains publications on theory,
applications, and design methods of Intelligent Systems and Intelligent Computing. Virtually
all disciplines such as engineering, natural sciences, computer and information science, ICT,
economics, business, e-commerce, environment, healthcare, life science are covered. The list
of topics spans all the areas of modern intelligent systems and computing.
The publications within “Advances in Intelligent Systems and Computing” are primarily
textbooks and proceedings of important conferences, symposia and congresses. They cover
significant recent developments in the field, both of a foundational and applicable character.
An important characteristic feature of the series is the short publication time and world-wide
distribution. This permits a rapid and broad dissemination of research results.
Advisory Board
Chairman
Nikhil R. Pal, Indian Statistical Institute, Kolkata, India
e-mail: nikhil@isical.ac.in
Members
Rafael Bello Perez, Universidad Central “Marta Abreu” de Las Villas, Santa Clara, Cuba
e-mail: rbellop@uclv.edu.cu
Emilio S. Corchado, University of Salamanca, Salamanca, Spain
e-mail: escorchado@usal.es
Hani Hagras, University of Essex, Colchester, UK
e-mail: hani@essex.ac.uk
László T. Kóczy, Széchenyi István University, Győr, Hungary
e-mail: koczy@sze.hu
Vladik Kreinovich, University of Texas at El Paso, El Paso, USA
e-mail: vladik@utep.edu
Chin-Teng Lin, National Chiao Tung University, Hsinchu, Taiwan
e-mail: ctlin@mail.nctu.edu.tw
Jie Lu, University of Technology, Sydney, Australia
e-mail: Jie.Lu@uts.edu.au
Patricia Melin, Tijuana Institute of Technology, Tijuana, Mexico
e-mail: epmelin@hafsamx.org
Nadia Nedjah, State University of Rio de Janeiro, Rio de Janeiro, Brazil
e-mail: nadia@eng.uerj.br
Ngoc Thanh Nguyen, Wroclaw University of Technology, Wroclaw, Poland
e-mail: Ngoc-Thanh.Nguyen@pwr.edu.pl
Jun Wang, The Chinese University of Hong Kong, Shatin, Hong Kong
e-mail: jwang@mae.cuhk.edu.hk
More information about this series at http://www.springer.com/series/11156

Marek Gzik Ewaryst Tkacz
•
Zbigniew Paszenda Ewa Piętka

•
Editors
Innovations in Biomedical
Engineering
123
Editors
Marek Gzik Zbigniew Paszenda
Faculty of Biomedical Engineering, Faculty of Biomedical Engineering,
Department of Biomechatronics Department of Biomaterials and Medical
Silesian University of Technology Devices Engineering
Zabrze Silesian University of Technology
Poland Zabrze
Poland
Ewaryst Tkacz
Faculty of Biomedical Engineering, Ewa Piętka
Department of Biosensors and Biomedical Faculty of Biomedical Engineering,
Signals Processing Department of Informatics and Medical
Silesian University of Technology Equipment
Zabrze Silesian University of Technology
Poland Zabrze
Poland
ISSN 2194-5357 ISSN 2194-5365 (electronic)

Advances in Intelligent Systems and Computing
ISBN 978-3-319-70062-5 ISBN 978-3-319-70063-2 (eBook)
https://doi.org/10.1007/978-3-319-70063-2
Library of Congress Control Number: 2017956134
© Springer International Publishing AG 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
In the period of last two decades it is not hard to observe unusual direct progress of
civilization in many fields concerning conditionality coming out from technical the-
ories or more generally technical sciences. We experience extraordinary dynamics
of the development of technological processes including different fields of daily life
which concerns particularly ways of communicating. We are aspiring for dissemi-
nating of the view that the success in the concrete action is a consequence of the
wisdom won over, collected and appropriately processed. They are talking straight
out about the coming into existence of the information society.
In such a context the meeting of the specialists dealing with the widely under-
stood innovations in Biomedical Engineering give a new dimension associated with
promoting something like the new quality. Because having the innovative approach
as a pointer in today’s world of changing attitudes and socio-economic conditions
can be perceived as one of the most important advantages. It results from the univer-
sal globalization letting observe oneself of surrounding world. Thanks to the devel-
opment of new biotechnologies comming out from the rapid progress in Biomedical
Sciences comprehending the contemporary needs of surrounding world it may be
said almost without any risk that life without biomedical sciences would stopped
existing.
At present, as it seems, implementing the universal standardization of the transfer
and the processing of information is the most important issue what in the significant
way influences for expanding the circle of Biomedical applications. It is a kind of
challenge to put the proper weight into particular branches covered by biomedi-
cal engineering and therefore we decided to edit the book as four part elaboration
covering biomaterials, biomechanics, biomedical informatics and last but not least
biomedical signals processing. One should aspire to it permanent integration rather
than the disintegration to progress in the context of the technological development.
Hence the constant observation and the appropriate problem analysis of biomedical
sciences as well as checking the technologies development and their applications is
picking the great importance up.
The monograph returned to hands of readers being a result of meeting specialists
dealing with above mentioned issues should in the significant way contribute to the
v
vi Preface
success in implementing consequences of human imagination into the social life. We

believe being aware of a human weakness and an imperfection that the monograph
presenting of a joint effort of the increasing numerically crowd of professionals and
enthusiasts will influence the further technology development regarding generaly
understood biomedicine with constantly expanding spectrum of its applications.
The last part of this preface will be devoted to express our great thanks and appre-
ciation to all the contributors of this book, which were listed in the special section as
"contributors list" and to persons who gave us an unusual help in final editing pro-
cess. Specials thanks are transfered to Szymon Sieciński for incredible engagement
and help in creating the final version of this book.
Zabrze, Ewaryst TKACZ

October 2017 Marek GZIK
Zbigniew PASZENDA
Ewa PIETKA
˛
Contents
Part I Informatics in medicine
1 Approach for spectrogram analysis in detection of selected

pronunciation pathologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Wojciech Bodusz, Zuzanna Miodońska and Paweł Badura
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Image processing and extraction of image features . . . . . . 6
1.2.3 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.4 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.1 Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 Automatic segmentation of lung cancer cells with the new
parameters by using methods of image processing and analysis. . . . . 13
Przemysław J˛edrusik, Robert Koprowski, Ilona Bednarek and Zygmunt
Wróbel
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Materials in experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
vii
viii Contents
3 Breast Cancer Segmentation Method

in Ultrasound Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Marta Galińska, Weronika Ogiegło, Agata Wijata, Jan Juszczyk and
Joanna Czajkowska
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4 Detection and Tracking of the Biopsy Needle Using Ultrasound
Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Agata Wijata, Żaneta Ranosz, Marta Galińska, Jan Juszczyk and
Joanna Czajkowska
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.2.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5 Detection of Wet Age-related Macular Degeneration in OCT
Images: A Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Anam Haq and Szymon Wilk
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.3.1 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.3.2 Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.3.3 Feature Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.3.4 Construction of Classifiers . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
6 Gender recognition using artificial neural networks and data
coming from force plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Jakub Krzysztof Grabski, Tomasz Walczak, Martyna Michałowska, and
Magdalena Cieślak
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.2 Human gait and ground reaction forces . . . . . . . . . . . . . . . . . . . . . . . 54
6.3 Data from the force plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.4 Artificial neural network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Contents ix
7 Human Sperm Morphology Analysis using a Digital Holographic

Microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Emil Fabian, Marzena Kamieniczna, Maciej Kurpisz and Ewa
Stachowska
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
7.2 Materials and apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
7.3 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
7.4 Measurements and results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
7.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
8 Idea and measurement methods used in bioimpedance spectroscopy. 69
Barbara Szuster, Zbigniew Szczurek, Dawid Roj, Pawel Kowalski,
Aleksander Sobotnicki and Jakub Woloszyn
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
8.2 Impedance measurements - physical basis . . . . . . . . . . . . . . . . . . . . . 70
8.3 Electrical bioimpedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
8.4 Body impedance device developed at ITAM . . . . . . . . . . . . . . . . . . . 72
8.5 Comparison of different commercially available devices for
bioimpedance measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
8.6 Different methods for the measurement and analysis of body
composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
8.6.1 DXA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
8.6.2 Antropometric measurements . . . . . . . . . . . . . . . . . . . . . . . 77
8.6.3 Analysis of neuronal activation . . . . . . . . . . . . . . . . . . . . . . 77
8.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
9 Prototype measurement system for spatial analysis of speech signal

for speech therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Kinga Kostera, Wojciech Wi˛eclawek, and Michał Kr˛ecichwost
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
9.2.1 Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.2.2 Measuring system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.2.3 Signal processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
9.3 Experiments and results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
9.3.1 Distance from sound source to microphone array . . . . . . . 82
9.3.2 Influence of recorded signals frequency . . . . . . . . . . . . . . . 84
9.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
9.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
x Contents
10 Shear Wave Elastography and Strain Elastography: a Study on a

Phantom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Bartłomiej Pyciński, Jan Juszczyk, and Rafał Obuchowicz
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
10.2.1 US phantom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
10.3 Experiments and results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
10.3.1 Shear Wave Elasticity Imaging . . . . . . . . . . . . . . . . . . . . . . 90
10.3.2 Strain Elastography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
10.3.3 Phantom ultrasound velocity . . . . . . . . . . . . . . . . . . . . . . . . 94
10.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
11 A simple setup for repeatability analysis of a low-level laser therapy
scanner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Dawid Kucharski and Jagoda Nowak
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
11.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
11.2.1 The Polaris 2 scanner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
11.2.2 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
11.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
11.3.1 The speed of the beam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
11.3.2 Laser beam geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
11.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
12 The influence of music genres on human emotionality . . . . . . . . . . . . . 107
Monika N. Bugdol, Marcin D. Bugdol and Tomasz Smreczak
12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
12.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
12.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Part II Signal analysis
13 2017 Monitoring and Teletransmission of Medical-Data in Heart

Failure. First Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Mateusz Tajstra, Piotr Rozentryt, Elżbieta Gadula-Gacek, Jacek
Niedziela, Elżbieta Adamowicz-Czoch, Aneta Ociessa, Adam Gacek,
Arkadiusz Gwóźdź, Marcin Wilczek, Aleksander Płaczek, Konrad
Wojciechowski, Adam Sokal, Zbigniew Kalarus, Mariusz Gasior, ˛ and
Lech Poloński
13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
13.2.1 Information Platform for Integrated Data Collection . . . . 119
Contents xi
13.2.2 EXTRABIOTEL-HF study . . . . . . . . . . . . . . . . . . . . . . . . . 120

13.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
13.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
14 Application of Discrete Cosine Transform for Pre-Filtering Signals
in Electrogastrography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Dariusz Komorowski and Barbara Mika
14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
14.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
14.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
14.4 Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
15 Determining Heart Rate Beat-to-Beat from Smartphone

Seismocardiograms: Preliminary Studies . . . . . . . . . . . . . . . . . . . . . . . 133
Szymon Sieciński and Paweł Kostka
15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
15.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
15.2.1 Experiment protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
15.2.2 HRV Estimation Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 135
15.2.3 Signal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
15.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
15.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
16 Field modelling of several electromagnetic drive variants of the
micropump blood transfusion device . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Sebastian Bartel, Zbigniew Pilch and Tomasz Trawi ´ nski
16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
16.2 Trends in developing ventricular assist devices and total artificial
hearts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
16.3 Geometrical modelling of electromagnetic pulsatile artificial heart 144
16.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
17 Minimizing interference from cardiac stimulator pulse in the
ECG recordings during the diagnostics of myocardial ischemia by
non-invasive transcutaneous cardiac stimulation . . . . . . . . . . . . . . . . . 151
Jerzy Gałecka, Fryderyk Prochaczek, Adam Gacek and Hanna
Winiarska-Prochaczek
17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
17.2 Sources of interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
17.3 Methods of interference suppression . . . . . . . . . . . . . . . . . . . . . . . . . 154
17.4 Minimizing interference in the precordial leads . . . . . . . . . . . . . . . . 157
17.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
xii Contents
18 Novel tumor protein markers collection by the use of highly porous

organic material for the upper and lower respiratory system –
preliminary results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Andrzej S. Swinarew, Barbara Mika, Jarosław Paluch Jadwiga Gabor,
Marta Ł˛eżniak, Hubert Okła, Tomasz Flak, Beata Swinarew, and
Klaudia Kubik
18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
18.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
18.2.1 Synthesis of sorbent material . . . . . . . . . . . . . . . . . . . . . . . . 164
18.2.2 MALDI-TOF analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
18.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
19 Statistical Analysis of the Impact of Molecular Descriptors on
Antimicrobial Activity of Thiourea Derivatives Incorporating
3-amino-1,2,4-triazole Scaffold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Anna Filipowska, Wojciech Filipowski, Ewaryst Tkacz, and Monika
Wujec
19.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
19.2 Experimental Works and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
19.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
19.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
20 The Face Tracking System for Rehabilitation Robotics Applications . 185
Paweł Raif and Ewaryst Tkacz
20.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
20.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
20.2.1 Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
20.2.2 Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
20.2.3 Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
20.2.4 Camera Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
20.2.5 Vision analysis and face tracing task . . . . . . . . . . . . . . . . . . 188
20.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
20.4 Conclusions and future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
21 The higher-order spectra as a tool for the identification of patients
diagnosed with various cardiac diseases . . . . . . . . . . . . . . . . . . . . . . . . . 193
Zbigniew Budzianowski, Ewaryst Tkacz, Wojciech Oleksy, Małgorzata
Garbacik
21.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
21.2 Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
21.2.1 A description of the data used . . . . . . . . . . . . . . . . . . . . . . . 194
21.2.2 Signal processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
21.3 Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Contents xiii
21.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

21.4.1 Bispectral analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
21.4.2 Bicohrent analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
21.4.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
21.4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
21.5 List of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
22 The prototype of wearable sensors system for supervision of patient
rehabilitation using artificial intelligence methods . . . . . . . . . . . . . . . . 205
Eliasz Kańtoch, Dominik Grochala, Marcin Kajor, and Dariusz
Kucharski
22.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
22.2 Architecture of the developed prototype . . . . . . . . . . . . . . . . . . . . . . 207
22.2.1 System hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
22.2.2 Multi-sensor acquisition unit . . . . . . . . . . . . . . . . . . . . . . . . 207
22.2.3 ECG acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
22.3 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
22.4 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
22.4.1 4.1 Dataset preparation and processing . . . . . . . . . . . . . . . . 210
22.4.2 4.2 Classification with machine learning algorithms. . . . . 211
22.6 Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Part III Modelling and simulations in biomechanics
23 Assessment of balance of older people living at a social welfare home 217

Katarzyna Jochymczyk-Woźniak, Katarzyna Nowakowska, Robert
Michnik, Agnieszka Nawrat-Szołtysik and Wioletta Górka
23.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
23.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
23.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
24 Assessment of locomotor functions of patients suffering from
cerebral palsy qualified to treat by different methods . . . . . . . . . . . . . 225
Katarzyna Jochymczyk-Woźniak, Katarzyna Nowakowska, Robert
Michnik, Agnieszka Konopelska, Jerzy Luszawski and Marek Mandera
24.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
24.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
24.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
24.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
xiv Contents
25 Body part accelerations evaluation for chosen techniques in martial

arts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Sebastian Glowiński, Andrzej Blażejewski and Tomasz Krzyzy ˙ ´ nski
25.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
25.2 Ukemi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
25.2.1 Koho Ukemi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
25.2.2 Yoko Ukemi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
25.2.3 Mae Ukemi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
25.2.4 Zenpo Kaiten Ukemi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
25.3.1 Participant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
25.3.2 Wireless sensing system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
25.3.3 Experiment procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
25.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
25.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
26 Determination of the number and frequency of the steps for gait
with elbow crutches based on a crutch acceleration . . . . . . . . . . . . . . . 245
Magdalena Długosz, Piotr Wodarski, Andrzej Bieniek, Miłosz Chrzan,
Marek Gzik, Kamil Joszko and Jarosław Derejczyk
26.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
26.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
26.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
26.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
26.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
27 Factors influencing on mechanical properties of porcine skin
obtained in tensile test-preliminary studies . . . . . . . . . . . . . . . . . . . . . . 255
Aneta Liber-Kneć and Sylwia Łagan
27.1 Intoduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
27.2 Material and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
27.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
28 Flow of subretinal liquid through the retinal hole after surgery –

mechanical model and FEM simulations . . . . . . . . . . . . . . . . . . . . . . . . 263
Tomasz Walczak, Paweł Fritzkowski, Marcin Stopa, and Martyna
Michałowska
28.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
28.2 Numerical simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
28.3 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Contents xv
29 Impact of Vessel Mechanical Properties on Hemodynamic

Parameters of Blood Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
Wojciech Wolański, Bożena Gzik-Zroska, Kamil Joszko, Edyta
Kawlewska, Marta Sobkowiak, Marek Gzik and Wojciech Kaspera
29.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
29.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
29.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
29.4 Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
30 Influence of strain rates on the hyperelastic material models
parameters of pig skin tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Sylwia Łagan and Aneta Liber-Kne ´ c
30.1 Intoduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
30.2 Material and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
30.3 Results and Discusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
30.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
31 Methodology of multicriterial optimization of geometric features of
an orthopedic implant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Małgorzata Muzalewska and Wojciech Moczulski
31.1 Origin of work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
31.1.1 Multicriterial optimization . . . . . . . . . . . . . . . . . . . . . . . . . . 290
31.1.2 Immune algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
31.2 Methodology of the optimal selection of geometric features of
an orthopedic implant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
31.2.1 Design assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
31.2.2 Criteria of concept evaluation - Strength . . . . . . . . . . . . . . 292
31.2.3 Criteria of concept evaluation - blood flow . . . . . . . . . . . . . 292
31.2.4 Objective function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
31.2.5 Pareto Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
31.3 Multicriterial optimization method using Immune Algorithm . . . . . 293
31.4 Verification and validation of the described methodology . . . . . . . . 294
31.4.1 Optimisation results by using the objective function . . . . . 294
31.4.2 Simulation results using the Pareto front . . . . . . . . . . . . . . 295
31.4.3 Strength verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
31.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
xvi Contents
32 Research on the stability of the users of chair with a spherical base . . 299
Robert Michnik, Miłosz Chrzan, Piotr Wodarski, Andrzej Bieniek,
Katarzyna Nowakowska, Anita Pollak and Andrzej Mitas
32.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
32.2 Research methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
32.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
32.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
32.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
Part IV Engineering of biomaterials
33 Assessment Of Hip Endoprosthesis Cups After Reimplantation . . . . . 311

Anita Kajzer, Ewelina Głab, ˛ Wojciech Kajzer, Tomasz Wróbel, and
Jacek Semenowicz
33.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
33.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
34 The heat treatment influence on the structure and mechanical
properties of Ti6Al4V alloy manufactured by SLM technology . . . . . 319
Marta Kiel-Jamrozik, Wojciech Jamrozik and Ilona Witkowska
34.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
34.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
34.3.1 Microstructure of Ti6Al4V produced by SLM . . . . . . . . . . 322
34.3.2 Results of the mechanical properties of SLM Ti6Al4V . . 323
34.3.3 Results of roughness measurements of Ti6Al4V alloy
fabricated by SLM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
34.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
35 The Funcionalization of Grade 4 Surface Used for Blood Contacting

Implants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Marcin Basiaga, Magdalena Antonowicz, Witold Walke, Zbigniew
Paszenda, and Bogusław Zi˛ebowicz
35.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
35.2.1 Preparation of samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
35.2.2 Electrochemical properties . . . . . . . . . . . . . . . . . . . . . . . . . . 331
35.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
35.3.1 Electrochemical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . 332
35.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
Contents xvii
36 The influence of implantation on mechanical degradation of the

nanotubular oxide layer on titanium screws . . . . . . . . . . . . . . . . . . . . . 339
Katarzyna Arkusz
36.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
36.2 Experimental details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
36.2.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
36.2.2 Fabrication of TNT on self-cutting cervical Ti6Al4V
ELI screw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
36.2.3 Electrochemical measurements . . . . . . . . . . . . . . . . . . . . . . 341
36.3.1 Surface morphology of anodized and non-modified
Ti6Al4V ELI screws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
36.3.2 The OCP evaluation of anodized and non-modified
Ti6Al4V ELI screws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
36.3.3 EIS measurements of Ti6Al4V ELI screws . . . . . . . . . . . . 344
36.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
37 The Influence of Technology on the Physicochemical and

Electrochemical Properties of the Prosthetic Materials . . . . . . . . . . . . 349
Anna Zi˛ebowicz, Anna Woźniak, and Bogusław Zi˛ebowicz
37.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
37.2 Material and Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
37.2.1 Surface roughness measurements . . . . . . . . . . . . . . . . . . . . 351
37.2.2 Wettability test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
37.2.3 Potentiodynamic test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
37.2.4 Microscope observation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
37.2.5 EIS test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
37.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
37.3.1 Surface roughness measurements . . . . . . . . . . . . . . . . . . . . 353
37.3.2 Wettability test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
37.3.3 Potentiodynamic test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
37.3.4 Microscope observation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
37.3.5 EIS tets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
37.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
38 Influence of Calcination Temperatures on the Morphology and
Specific Surface Area Development of ZrO2 -CeO2 -Y2 O3 -Al2 O3
Powders Obtained via Sol-Gel Route . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
Damian S. Nakonieczny, Zbigniew K. Paszenda, Justyna Majewska,
Sabina Drewniak, Wojciech Bogacz, Magdalena Antonowicz, Agata
Sambok, and Cezary Krawczyk
38.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
38.2.1 Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
xviii Contents
38.2.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

38.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
38.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
38.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
39 The method of fretting wear assessment with the application of 3D
laser measuring microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Jarosław Sidun, Jan Ryszard Dabrowski ˛
39.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
39.2 Research material and methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 371
39.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
39.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
40 The cell niches reproducing surface structure . . . . . . . . . . . . . . . . . . . . 379
Roman Major, Klaudia Trembecka-Wojciga, Jürgen Markus Lackner,
and Bogusław Major
40.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
40.2 Materials and diagnostic methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
40.2.1 Hydrogel coating deposition . . . . . . . . . . . . . . . . . . . . . . . . 380
40.2.2 Surface of modification of the controlled residual stresses 381
40.2.3 In vitro analysis of blood-matreial interaction -
Impact-R test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
40.2.4 Methods of microstructure analysis . . . . . . . . . . . . . . . . . . . 383
40.3.1 Blood-material interaction . . . . . . . . . . . . . . . . . . . . . . . . . . 385
40.3.2 Endothelialisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
40.4 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
41 Efect Of Sterylization And Long-Term Exposure To Artificial

Urine On Corrosion Behaviour Of Metallic Biomaterials With
Poly(Glikolide-Co-Kaprolactone) Coatings . . . . . . . . . . . . . . . . . . . . . . 391
Wojciech Kajzer, Joanna Jaworska, Katarzyna Jelonek, Janusz
Szewczenko, Katarzyna Nowińska, and Anita Kajzer
41.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
41.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
42 Mechanical characterization of biodegradable materials used in
surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Angela Andrzejewska
42.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Contents xix

42.2.1 Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
42.2.2 Mass change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
42.2.3 Mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
42.2.4 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
42.3.1 Test of mass change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
42.3.2 Test of static properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
42.3.3 Comparison between mass change and mechanical
properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
42.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409

List of Contributors
Elżbieta Adamowicz-Czoch
Third Department of Cardiology, SMDZ in Zabrze, Medical University of Silesia
in Katowice, Silesian Centre for Heart Disease, Zabrze, Poland
Angela Andrzejewska
UTP University of Science and Technology in Bydgoszcz, Al. Prof. S. Kaliskiego 7,
85-796 Bydgoszcz, POLAND
e-mail: angela.andrzejewska@utp.edu.pl
Magdalena Antonowicz
Department of Biomaterials and Medical Devices Engineering, Faculty of
Biomedical Engineering, Silesian University of Technology, Roosevelta 40 st.,
Zabrze, Poland
Katarzyna Arkusz
Biomedical Engineering Division, Department of Mechanical Engineering,
University of Zielona Gora, Zielona Gora, Poland
e-mail: k.arkusz@ibem.uz.zgora.pl
Paweł Badura
Faculty of Biomedical Engineering, Silesian University of Technology,
Roosevelta 40, 41-800 Zabrze, Poland
Sebastian Bartel
Silesian University of Technology, Faculty of Electrical Enginieering,
Mechatronics Department
Akademicka str. 2A, 44-100 Gliwice, Poland,
e-mail: sebastian.bartel@polsl.pl
Marcin Basiaga
Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze,
Poland
e-mail: Marcin.Basiaga@polsl.pl
xxi
xxii List of Contributors
Ilona Bednarek
Department of Biotechnology and Genetic Engineering, Medical University of
Silesia,
School of Pharmacy with the Division of Laboratory Medicine in Sosnowiec,
ul. Jedności 8, 41-200 Sosnowiec, Poland
e-mail: ibednarek@sum.edu.pl
Andrzej Bieniek
Silesian University of Technology, faculty of Biomedical Engineering, Poland
Andrzej Błażejewski
Technical University of Koszalin, Sniadeckich 2, 75-453 Koszalin, Poland
Wojciech Bodusz
e-mail: wojciech.bodusz93@gmai.com
Wojciech Bogacz
Department of Chemical Engineering and Process Design, Faculty of Chemistry,
Silesian University of Technology, M. Strzody 7 st., Gliwice, Poland
Zbigniew Budzianowski
Silesian University of Technology, 41-800 Zabrze, Poland,
e-mail: zbigniew.budzianowski@polsl.pl
Marcin D. Bugdol
Faculty of Biomedical Engineering, Department of Informatics and Medical
Equipment, Silesian University of Technology, Zabrze, Poland
Monika N. Bugdol
e-mail: monika.bugdol@polsl.pl
Miłosz Chrzan
Magdalena Cieślak
Institute of Applied Mechanics, Faculty of Mechanical Engineering and Manage-
ment, Poznan University of Technology, Piotrowo 3, 60-965 Poznań, Poland
Joanna Czajkowska
Silesian University of Technology, Faculty of Biomedical Engineering, Zabrze,
Poland
Jan Ryszard Dabrowski
˛
Bialystok University of Technology, Mechanical Faculty,
Wiejska 45c, 15-351 Białystok
e-mail: j.dabrowski@pb.edu.pl
Jarosław Derejczyk
List of Contributors xxiii
Katowice - Szpital Geriatryczny im. Jana Pawła II

Magdalena Długosz
e-mail: magdadlugosz93@gmail.com
Sabina Drewniak
Department of Optoelectronics, Silesian University of Technology, B. Krzy-
woustego 2 st., Gliwice, Poland
Emil Fabian
Department of Metrology and Measurement Systems, Institute of Mechanical
Technology, Faculty of Mechanical Engineering and Management, Poznan
University of Technology, ul. Piotrowo 3, PL60-965 Poznań, Poland
Anna Filipowska
Faculty of Automatic Control, Electronics and Computer Science, Silesian
University of Technology, Poland
Department of Biosensors and Processing of Biomedical Signals, Silesian
e-mail: afilipowska@polsl.pl
Wojciech Filipowski
Faculty of Automatic Control, Electronics and Computer Science, Silesian
Tomasz Flak
Institute of Materials Science, Faculty of Computer Science and Material Science,
University of Silesia in Katowice,
75 Pułku Piechoty 1A, 41-500 Chorzów, Poland
Paweł Fritzkowski
Faculty of Mechanical Engineering and Management,
Poznań University of Technology
Jana Pawła II 24, 60-965 Poznań, Poland
Jadwiga Gabor
Adam Gacek
Institute of Medical Technology and Equipment, Zabrze, Poland
Marta Galińska
Silesian University of Technology, Faculty of Automatic Control, Electronics
and Computer Science, Gliwice, Poland
e-mail: marta.galinska@polsl.pl
Jerzy Gałecka
Institute of Medical Technology and Equipment ITAM, Zabrze, Poland
xxiv List of Contributors
e-mail: jerzy.galecka@itam.zabrze.pl
Małgorzata Garbacik
Silesian University of Technology, 41-800 Zabrze, Poland
Mariusz Gasior
˛
Ewelina Głab
˛
Biomedical Engineering, Silesian University of Technology, ul. Roosevelta 40,
44-800 Zabrze, Poland
Sebastian Głowiński
Technical University of Koszalin, Sniadeckich 2, 75-453 Koszalin, Poland,
e-mail: sebastian.glowinski@tu.koszalin.pl
Wioletta Górka
Students Scientific ”BIOKREATYWNI”, Faculty of Biomedical Engeenering,
Silesian Univesity of Technology, Poland
Jakub Krzysztof Grabski
e-mail: jakub.grabski@put.poznan.pl
Dominik Grochala
AGH University of Science and Technology, 30 Mickiewicza Av., 30-059 Kraków,
Poland
Arkadiusz Gwóźdź
WASKO SA, Gliwice, Poland
Marek Gzik
Department of Biomechatronics, Faculty of Biomedical Engineering,
Silesian University of Technology, ul. F. D. Roosevelta 40, 41-800 Zabrze, Poland
e-mail: marek.gzik@polsl.pl
Anam Haq
Poznan University of Technology, Poznan, Poland
e-mail: anam.haq@put.poznan.pl
Wojciech Jamrozik
Faculty of Mechanical Engineering, Silesian University of Technology,
Konarskiego str 18a, 44-100 Gliwice, Poland
Joanna Jaworska
Centre of Polymer and Carbon Materials of the Polish Academy of Sciences
Zabrze, Poland
Katarzyna Jelonek
List of Contributors xxv
Centre of Polymer and Carbon Materials of the Polish Academy of Sciences

Zabrze, Poland
Przemysław J˛edrusik
Department of Computer Biomedical Systems, University of Silesia,
Institute of Computer Science,
ul. B˛edzińska 39, 41âĂŘ200 Sosnowiec, Poland
e-mail: pjedrusik@o2.pl
Katarzyna Jochymczyk-Woźniak
Department of Biomechatronics, Faculty of Biomedical Engeenering, Silesian
Univesity of Technology, Poland
e-mail: Katarzyna.Jochymczyk-Wozniak@polsl.pl
Kamil Joszko
Jan Juszczyk
Marcin Kajor
Poland
Anita Kajzer
e-mail: anita.kajzer@polsl.pl
Wojciech Kajzer
e-mail: wojciech.kajzer@polsl.pl
Zbigniew Kalarus
Department of Cardiology, Congenital Heart Diseases and Electrotherapy, Silesian
Center for Heart Diseases, Medical University of Silesia, Zabrze, Poland
Marzena Kamieniczna
Institute of Human Genetics Polish Academy of Science,
ul. Strzeszyńska 32, PL60-479 Poznań, Poland
Eliasz Kańtoch
Poland
e-mail: kantoch@agh.edu.pl
Edyta Kawlewska
xxvi List of Contributors

Silesian University of Technology, ul. F. D. Roosevelta 40, 41-800 Zabrze, Poland
e-mail: edyta.kawlewska@polsl.pl
Marta Kiel-Jamrozik
Roosvelta str 40, 41-800 Zabrze, Poland
e-mail: marta.kiel-jamrozik@polsl.pl
Dariusz Komorowski
Silesian University of Technology, Faculty of Biomedical Engineering, Department
of Biosensors and Processing of Biomedical Signals, Zabrze, Poland
e-mail: dariusz.komorowski@polsl.pl
Agnieszka Konopelska
Institute of Theoretical and Applied Mechanics, the Silesian University of
Technology, Konarskiego 18a, 44-100 Gliwice, Poland
Robert Koprowski
ul. B˛edzińska 39, 41-200 Sosnowiec, Poland
e-mail: robert.koprowski@us.edu.pl
Kinga Kostera
Faculty of Biomedical Engineering, Silesian University of Technology, Roosevelta
40, 41-800 Zabrze, Poland
Paweł Kostka
Silesian University of Technology, Department of Biosensors and Biomedical
Signals Processing, Zabrze, Poland
e-mail: pkostka@polsl.pl
Cezary Krawczyk
Department of Dental Technology, Medical College of Zabrze, 3 maja 63 st.,
Zabrze, Poland
Michał Kr˛ecichwost
Tomasz Krzyżyński
Technical University of Koszalin, Sniadeckich 2, 75-453 Koszalin, Poland
Klaudia Kubik
Dariusz Kucharski
AGH University of Science and Technology,
List of Contributors xxvii
30 Mickiewicza Av., 30-059 Kraków, Poland

Dawid Kucharski
Poznan University of Technology,
Faculty of Mechanical Engineering and Management
Institute of Mechanical Technology
Division of Metrology and Measurement Systems
Poznan ul. Piotrowo 3, 60-965, Poland
e-mail: dawid.kucharski@put.poznan.pl
Maciej Kurpisz
Institute of Human Genetics Polish Academy of Science, ul. Strzeszyńska 32,
PL60-479 Poznań, Poland
Jürgen Markus Lackner
Joanneum Research Forschungsges mbH, Institute of Surface Technologies and
Photonics, Functional Surfaces, Leobner Strasse 94, A-8712 Niklasdorf, Austria
Aneta Liber-Kneć
Cracow University of Technology, Warszawska 24, 31-155 Cracow, Poland
e-mail: aliber@pk.edu.pl
Jerzy Luszawski
Upper Silesian Children’s Health Centre in Katowice, Medyków 16, 40-751
Katowice, Poland
Sylwia Łagan
Cracow University of Technology, Warszawska 24, 31-155 Cracow, Poland
e-mail: slagan@mech.pk.edu.pl
Marta Ł˛eżniak
Justyna Majewska
Department of Biosensors and Biomedical Signal Processing, Faculty of Biomedi-
cal Engineering, Silesian University of Technology,
Roosevelt 40 st., 41-800 Zabrze, Poland
Bogusław Major
Institute of Metallurgy and Materials Science; Polish Academy of Sciences,
30-059 Cracow, Reymonta St.25, Poland
Roman Major
e-mail: r.major@imim.pl
Marek Mandera
Upper Silesian Children’s Health Centre in Katowice,
xxviii List of Contributors
Medyków 16, 40-751 Katowice, Poland

Martyna Michałowska
Robert Michnik
e-mail: Robert.Michnik@polsl.pl
Barbara Mika
of Biosensors and Processing of Biomedical Signals, Zabrze, Poland
Zuzanna Miodońska
Andrzej Mitas
Department of Informatics and Medical Equipment, Faculty of Biomedical
Engeenering, Silesian Univesity of Technology, Poland
Wojciech Moczulski
Silesian University of Technology
e-mail: wojciech.moczulski@polsl.pl
Małgorzata Muzalewska
e-mail: malgorzata.muzalewska@polsl.pl
Damian S. Nakonieczny
Biomedical Engineering, Silesian University of Technology, Roosevelta 40 st.,
Zabrze, Poland
e-mail: Damian.Nakonieczny@polsl.pl
Agnieszka Nawrat-Szołtysik
Chair of Nervous and Locomotor System Physiotherapy, Academy of Physical
Education in Katowice, Poland
Jacek Niedziela
Jagoda Nowak
Poznan ul. Piotrowo 3, 60-965, Poland
List of Contributors xxix
Katarzyna Nowakowska
Katarzyna Nowińska
Institute of Applied Geology, Faculty of Mining and Geology, Silesian University
of Technology, Gliwice, Poland.
Rafał Obuchowicz
Department of Radiology, Collegium Medicum,
Jagiellonian University, Cracow, Poland
Aneta Ociessa
Weronika Ogiegło
Poland
Hubert Okła
Wojciech Oleksy
Silesian University of Technology, 41-800 Zabrze, Poland
Jarosław Paluch
Department and Clinic of Laryngology, School of Medicine in Katowice,
Medical University of Silesia in Katowice,
Francuska 20/24, 40-027 Katowice, Poland
Zbigniew K. Paszenda
Poland
Ewa Pi˛etka
Zbigniew Pilch
Akademicka str. 2A, 44-100 Gliwice, Poland
e-mail: zbigniew.pilch@polsl.pl
Aleksander Płaczek
Anita Pollak
xxx List of Contributors
Institute of Psychology, University of Silesia, Katowice, Poland

Lech Poloński
Fryderyk Prochaczek
Institute of Medical Technology and Equipment ITAM, Zabrze, Poland
Bartłomiej Pyciński
e-mail: bartlomiej.pycinski@polsl.pl
Paweł Raif
Department of Biosensors and Processing of Biomedical Signals,
Biomedical Engineering Faculty at the Silesian University of Technology,
Zabrze, Poland
Żaneta Ranosz
of Informatics and Medical Equipment, Gliwice, Poland
Piotr Rozentryt
Agata Sambok
Biomedical Engineering, Silesian University of Technology,
Roosevelta 40 st., Zabrze, Poland
Jacek Semenowicz
Dr Janusz Daab Independent Public Regional Hospital of Trauma Surgery in
Piekary Slaskie, 62 Bytomska Street 41 - 940 Piekary Ślaskie,
˛ Polska
Jarosław Sidun
Bialystok University of Technology, Mechanical Faculty,
Wiejska 45c, 15-351 Białystok
e-mail: j.sidun@pb.edu.pl
Szymon Sieciński
Silesian University of Technology, Department of Biosensors and Biomedical
Signals Processing, Zabrze, Poland,
Tomasz Smreczak
Marta Sobkowiak
List of Contributors xxxi
Silesian University of Technology, ul. F. D. Roosevelta 40, 41-800 Zabrze, Poland,

e-mail: marta.sobkowiak@polsl.pl
Adam Sokal
Department of Cardiology, Congenital Heart Diseases and Electrotherapy, Silesian
Center for Heart Diseases, Medical University of Silesia, Zabrze, Poland
Aleksander Sobotnicki
ITAM, Roosevelta 118, Zabrze, Poland
Ewa Stachowska
Department of Metrology and Measurement Systems, Institute of Mechanical
Technology, Faculty of Mechanical Engineering and Management, Poznan
University of Technology, ul. Piotrowo 3, PL60-965 Poznań, Poland
e-mail: ewa.stachowska@put.poznan.pl
Marcin Stopa
Clinical Eye Unit and Pediatric Ophthalmology Service,
Heliodor Świ˛ecicki University Hospital, Poznań University of Medical Sciences,
Przybyszewskiego 49, 60-355, Poznań, Poland
Andrzej S. Swinarew
e-mail: andrzej.swinarew@us.edu.pl
Beata Swinarew
Institute for Engineering of Polymer Materials and Dyes,
Paint and Plastics Department,
44-100 Gliwice, Poland
Janusz Szewczenko
Zbigniew Szczurek
Barbara Szuster
e-mail: basiasz@itam.zabrze.pl
Mateusz Tajstra
Ewaryst Tkacz
Department of Biosensors and Processing of Biomedical Signals, Silesian
xxxii List of Contributors
Tomasz Trawiński
e-mail: tomasz.trawinski@polsl.pl
Klaudia Trembecka-Wójciga
Tomasz Walczak
e-mail: tomasz.walczak@put.poznan.pl
Witold Walke
Poland
Wojciech Wi˛ecławek
Agata Wijata
Silesian University of Technology, Faculty of Automatic Control, Electronics
and Computer Science, Gliwice, Poland
Poland
e-mail: agata.wijata@polsl.pl
Marcin Wilczek
e-mail: M.Wilczek@wasko.pl
Szymon Wilk
Poznan University of Technology, Poznan, Poland e-mail: szymon.wilk@cs.
put.poznan.pl
Hanna Winiarska-Prochaczek
Cardiologic Outpatient Clinic CARDIOTEST, Tychy, Poland
Ilona Witkowska
Roosvelta str 40, 41-800 Zabrze, Poland
Piotr Wodarski
e-mail: piotr.wodarski@polsl.pl
Konrad Wojciechowski
Polish Japanese Academy of Information Technology, Warszawa, Poland
List of Contributors xxxiii
Jakub Woloszyn
ITAM, Roosevelta 118, Zabrze, Poland,
Tomasz Wróbel
Department of Foundry, Faculty of Mechanical Engineering, Silesian University of
Technology, Towarowa 7, 44-100 Gliwice, Poland.
Zygmunt Wróbel
ul. B˛edzińska 39, 41-200 Sosnowiec, Poland
Monika Wujec
Department of Organic Chemistry, Medical University of Lublin, Poland
Anna Woźniak
Department of Biomaterials and Medical Devices Engineering, Silesian University
of Technology, Roosevelta 40, 41-800 Zabrze Poland,
Anna Zi˛ebowicz
Department of Biomaterials and Medical Devices Engineering, Silesian University
of Technology, Roosevelta 40, 41-800 Zabrze Poland
e-mail: anna.ziebowicz@polsl.pl
Bogusław Zi˛ebowicz
Department of Biomedical Materials Engineering, Silesian University of
Technology, Konarskiego 18a, 44-100 Gliwice Poland
e-mail: boguslaw.ziebowicz@polsl.pl
Part I
Informatics in medicine
Chapter 1
Approach for spectrogram analysis in detection
of selected pronunciation pathologies
Wojciech Bodusz1 , Zuzanna Miodońska1 , and Paweł Badura1
Abstract. An attempt to automatise selected pronunciation pathology detection in

preschool children is described in this paper. Consonant [Z] in various phonetic sur-
roundings is taken into consideration as eventual sigmatism indicator. The analysis
involves spectrogram analysis in terms of image processing methods used for feature
extraction and classification. Five dedicated features are defined and extracted, i.a.,
from a frequency sub-band of [1500, 6500] Hz. Binary classification using support
vector machine enables pathology detection. The system performance is evaluated
using sensitivity, specificity, and accuracy metrics in two cross-validation experi-
ments over a database of 140 speech recordings with 50 normative and 90 patho-
logical cases. Repeatable efficiency metrics at a ca. 85% level confirm the method
capabilities and encourage to develop the system for the speech diagnosis support.
Keywords: speech pathology, spectrogram analysis, image processing
1.1 Introduction
Speech is the major and most commonly used mean of communication [1, 2]. As an
acoustic wave, it covers the frequency range of 250-8000 Hz, yet the differences in
construction of speech organs cause to vary these values in individual cases. Speech
dysfunctions are widespread, especially among children. Thus, there is a need to
analyse and assess speech for diagnosis and therapy purposes [3, 4]. In many cases
speech signal is presented as a spectrogram (STFT – short-time Fourier transform),
which can be treated as an image [5, 6, 7]. It is possible to extract a number of
features from such an image, providing information about potential pronunciation
pathology [8]. The feature values can be obtained by using a variety of image pro-

e-mail: wojciech.bodusz93@gmail.com
© Springer International Publishing AG 2018 3

M. Gzik et al. (eds.), Innovations in Biomedical Engineering,
Advances in Intelligent Systems and Computing 623,
https://doi.org/10.1007/978-3-319-70063-2_1
4 Wojciech Bodusz et al.
cessing methods, such as filtration, thresholding or edge detection algorithms [9].

Extracted image features can be used to train and feed the classifier in order to
detect possible speech disorders. Systems for speech pathology detection can op-
erate directly on acoustic signals by means of, e.g., wavelet-based feature extrac-
tion [10, 11], involve STFT and spectral analysis [12, 13], or employ features dedi-
cated to speech analysis, e.g., mel-frequency cepstral coefficients (MFCC) [14].
The purpose of the study was to design a system for speech pathology detection
based on spectrogram analysis in terms of image processing techniques. A database
of 140 reference recordings was employed to perform the analysis, define the feature
vector, train, and validate the support vector machine (SVM) classifier. The record-
ings contain single realizations of Polish words with a consonant [Z] performed by
preschool children, assessed by the speech therapy expert for sigmatism annota-
tion (50 correct and 90 pathological pronunciations). The novelty of the study can
be identified in the idea of employing image processing techniques over the spec-
trogram of the audio signal in order to classify different pronunciation types. That
concerns, inter alia, definition and verification of original features and extraction
procedures.
The paper is organized as follows. Section 1.2 presents the STFT-image process-
ing methods, features definition, and classifier specification. Conducted experiments
and quantitative evaluation of the method are described in Section 1.3. The conclu-
sions are presented in Section 1.4.
1.2 Materials and methods
Fig. 1.1: General processing workow.
The general processing workow is presented in Fig. 1.1. Its subsequent stages
are described in following sections.
1 Image processing of speech spectrograms 5
1.2.1 Preprocessing
First, each recording is limited to the articulation of phoneme [Z]. Figs. 2(a) and 2(b)
present the time signal of an exemplary recording and the segmented phoneme Z, re-
spectively. Then, the STFT of the acoustic signal acquired with sampling frequency
fs of 44100 Hz and 16-bit resolution is determined with the following parameters:
window width at 0.01 · fs , time overlap half at of the window width, and the fre-
quency range of [100, 13000] Hz. The latter was adjusted to the analysed phoneme
[Z] effective range. Figs. 2(c) and 2(d) show spectrograms related to speech signals
(a) (b)
(c) (d)
Fig. 1.2: Time signal (top) and spectrogram (bottom) of the recording of preschool
child pronouncing a Polish word „róża” [ruZa] – eng. rose (a), (c) and with a seg-
mented Z phoneme only (b), (d). Spectrograms illustrated with a jet colormap.
from in Figs. 2(a) and 2(b), respectively.

(a) (b) (c)
Fig. 1.3: Example intermediate spectrogram analysis results: segment of raw spec-
trogram containing phoneme [Z] (a), after median filtering (b), and with detected
columns’ centroids (c).
1.2.2 Image processing and extraction of image features
Further analysis involves processing of spectrogram as a monochromatic image.

First, median filtering (with 5×5 kernel) is employed for noise reduction (Fig. 3(a)).
Next steps are directly related to extraction of image features. 5 features have been
investigated and selected for the classification and detection: image centroid, aver-
age centroid crossing ratio, entropy, standard deviation, and average gray level.
1.2.2.1 Image centroid
In order to calculate the location of image centroid C, the centroids of every column
of an image (Figs. 3(b), 1.4) are calculated using following formula:
N
∑ i · Ji, j
i=1
Cj = N
, (1.1)
∑ Ji, j
i=1
where i, j denote the row and column indices, Ji, j – a gray level of a pixel, and N
is the number of rows. The image centroid C is determined as an average centroid
over all image columns (Fig. 1.4).
Fig. 1.4: Example locations of column centroids (blue line) with the image centroid
C (red line) and average centroid crossing occurrences circled.
1.2.2.2 Average centroid crossing ratio (ACCR)
The number of intersections of the individual column centroid C j chart and the aver-
age image centroid C divided by the number of image column determines the value
of the average centroid crossing ratio ACCR (Fig. 1.4).
The other 3 features are extracted from a single column of the spectrogram, ob-
tained by averaging gray levels over all columns within the segment. Such columns
of normative (50) and pathological (90) cases combined together are presented in
Figs. 5(a) and Fig. 5(b), respectively. Clearly, the biggest differences between both
groups can be noticed in a [1500, 6500] Hz band, which leads to a closer look during
extraction of remaining features.
1.2.2.3 Entropy
The entropy E is calculated within the [1500, 6500] Hz band of the average column
using:
L
E = − ∑ p (li ) · log2 p (li ) , (1.2)
i=1
(a) (b)
Fig. 1.5: Combined average columns of all normative (a) and pathological (b)
recordings.
where li denotes the ith succeeding gray level appearing in the image with a fre-
quency p (li ) and L is the total number of gray levels within the image.
1.2.2.4 Average gray level
The sub-band column average gray level J¯ is determined within the [1500, 6500] Hz
band according to:
Nsb
∑ Ji
i=1
J¯ = (1.3)
Nsb
where Ji denotes the gray level of the ith of Nsb elements of the sub-band column.
1.2.2.5 Standard deviation
Similarly, the sub-band column gray level standard deviation SD is determined as:
Nsb
∑ (Ji − J)
¯
i=1
SD = , (1.4)
Nsb − 1
1.2.3 Classification
Support Vector Machine (SVM) [15] is used to classify recordings of speech signal
into one of two groups: norm or pathology. Radial basis function is used as the
kernel function of SVM.
1.2.4 Validation
In order to validate the algorithm, a database of 140 recordings was employed, con-
taining speech signal from preschool children. Each recording was assessed by a
speech therapy expert and annotated as normative (50 cases) or pathological (90)
in terms of sigmatism diagnosis. Two cross-validation methods were employed
for the detection verification: 4-fold cross-validation and the leave-one-out cross-
validation [16]. Verification of the algorithm relied on 3 detection efficiency metrics
(sensitivity, specificity, accuracy):
TP
sensitivity = , (1.5)
T P + FN
TN
speci f icity = , (1.6)
T N + FP
TP+TN
accuracy = , (1.7)
T P + FP + T N + FN
where T P, T N, FP, FN denote the number of true positive, true negative, false
positive, and false negative pathology detections, respectively.
1.3 Results and discussion
1.3.1 Experimental results
Tables 1.1 and 1.2 present results obtained for algorithm validation using both cross-
validation schemes: the 4-fold cross-validation (run 40 times, Table 1.1) with the
average, minimum, and maximum values of efficiency metrics and the leave-one-
out cross-validation (Table 1.2).
Table 1.1: Algorithm performance in the 4-fold cross-validation.
Metrics Accuracy Sensitivity Speci f icity

Maximum [%] 89.9 91.1 84.0
minimum [%] 82.1 86.7 76.0
mean [%] 85.9 88.4 80.3
Standard deviation [%] 1.4 1.2 1.3
Table 1.2: Algorithm performance in the leave-one-out cross-validation.
Accuracy [%] Sensitivity [%] Speci f icity [%]

85.7 88.9 82.0
1.3.2 Discussion
In both cross-validation schemes obtained results vary around 85%. In case of the
4-fold cross-validation, dispersion of results does not exceed 8 percentage points,
proving relatively high repeatability and reliability of classification. Results ob-
tained in the leave-one-out cross-validation (accuracy at 85,7%) confirm the ef-
ficiency of the method. Such detection accuracy indicators might be considered
acceptable in computer-aided speech diagnosis, therefore the proposed algorithm
shows promising capabilities for application and further development.
Possible problems leading to misclassification occurrences can be identified in
several sources, i.a., disproportion between number of recordings with normative
and pathological speech signal or diversified severity and types of speech patholo-
gies contained in the speech database. The latter can point out the direction for
further experiments within domain, i.e. a multi-class view at the speech pathology
issues in terms of various sigmatism diagnoses.
Since the study was focused strictly on detection of pathological phoneme real-
izations, detailed analysis of acquisition accuracy was not performed.
1.4 Conclusion
The aim of this study was to create a tool which would be able to perform the anal-
ysis of speech spectrograms. The processing involved extraction of image features
related to normative and pathological speech. The obtained results indicate, that the
goal has been reached at an acceptable level. Five significant features were defined
and selected for classification as an effect of thorough research. Most of the analysis
was focused on the frequency sub-band of [1500, 6500] Hz, identified as a distinc-
tive field in pronunciation classification. It should be noted, that way greater amount
of energy is cumulated in this band in case of normative speech.
In the future, other potentially problematic phonemes could be analysed. In order
to reach higher accuracy, the feature vector could be enlarged. Creating an efficient
tool which could detect pathologies of speech would be a great achievement, which
could help in speech diagnosis.
Acknowledgements
This research was supported by the Polish Ministry of Science and Silesian Univer-
sity of Technology statutory financial support No. BK-200/RIB1/2017.
References
1. R. Tadeusiewicz. Sygnał mowy. Wydawnictwa Komunikacji i Łaczności,

˛ Warszawa, 1988 (in
Polish).
2. A. Kaczmarek. Analiza sygnału mowy w foniatrii. Oddział Gdański PTETiS, 2006 (in Polish).
3. Z. Miodonska, M. D. Bugdol, and M. Krecichwost. Dynamic time warping in phoneme mod-
eling for fast pronunciation error detection. Computers in Biology and Medicine, 69:277–285,
2016.
4. M. Krecichwost, Z. Miodonska, J. Trzaskalik, J. Pyttel, and D. Spinczyk. Acoustic Mask for
Air Flow Distribution Analysis in Speech Therapy. In Information Technologies in Medicine,
ITIB 2016, vol. 1, volume 471 of Advances in Intelligent Systems and Computing, pages 377–
387, 2016.
5. T. Lampert and S. O’Keefe. A survey of spectrogram track detection algorithms. Department
of Computer Science, University of York, 2008.
6. T. Lampert and S. O’Keefe. On the detection of tracks in spectrogram images. Department of
Computer Science, University of York, 2012.
7. P.J. Durka. Mi˛edzy czasem a cz˛estoscia: ˛ elementy współczesnej analizy sygnałów. 1999 (in
Polish).
8. B. Pinkowski. Principal component analysis of speech spectrogram images. Computer Scie-
nce Department, Western Michigan University, 1996.
9. R. Gonzalez and R. Woods. Digital Image Processing. Prentice Hall, Upper Saddle River, NJ,
2002.
10. O. Farooq and S. Datta. Mel Filter-Like Admissible Wavelet Packet Structure for Speech
Recognition. IEEE Signal Processing Letters, vol. 8, No. 7, July 2001.
11. C.J. Long and S. Datta. Wavelet based feature extraction for phoneme recognition. Spoken
Language, 1996. ICSLP 96. Proceedings., Fourth International Conference, 1996.
12. L. Deng, G. Hinton, and B. Kingsbury. New types of deep neural network learning for speech
recognition and related applications: an overview. In 2013 IEEE International Conference on
Acoustics, Speech and Signal Processing. IEEE, May 2013.
13. A. Waibel, T. Hanazawa, G. Hinton, K. Shikano, and K. J. Lang. Phoneme recognition using
time-delay neural networks. IEEE Transactions on Acoustics, Speech, and Signal Processing,
37(3):328–339, Mar 1989.
14. Z. Miodonska, M. Krecichwost, and A. Szymanska. Computer-Aided Evaluation of Sibilants
in Preschool Children Sigmatism Diagnosis. In Information Technologies in Medicine, ITIB
2016, vol. 1, volume 471 of Advances in Intelligent Systems and Computing, pages 367–376,
2016.
15. C. Cortes and V. Vapnik. Support-vector networks. Machine Learning, 20(3):273–297, 1995.
16. S. Arlot and A. Celisse. A survey of cross-validation procedures for model selection. Statistics
Surveys, 4:40–79, 2010.
Chapter 2
Automatic segmentation of lung cancer cells
with the new parameters by using methods of
image processing and analysis.
Przemysław J˛edrusik1 , Robert Koprowski1 , Ilona Bednarek2 , and Zygmunt

Wróbel1
Abstract. Modern diagnostic methods allow to get multiple information regarding

research material. This work focused on the development of an algorithm for au-
tomatically determining the correct number of cells. The developed tool allows the
detection of cells as individual objects, searching for the objects significantly larger
than the sought and checking if they were a combination of objects. The algorithm
was based on additional parameters designated in its subsequent steps as well as
their respective correcting claimed searched result. Analyzed a large number of im-
ages, it was found that there is a close relationship between the surface area of the
cells, the degree of extension and the location and correct detection of objects that
are neither a cluster of cells, and nothing significant image artifacts. The developed
algorithm was written using Matlab software.
Keywords: image processing, algorithms, automatic, lung cancer, cell culture, cell
migration
2.1 Introduction
The huge number of medical information obtained in the process of computer - as-
sisted acquisition, allows to reach the growing number of different cases. Existing
barriers to obtaining large amounts of data, perform complex and time-consuming

ul. B˛edzińska 39, 41-200 Sosnowiec, Poland,
e-mail: pjedrusik@o2.pl,robert.koprowski@us.edu.pl · Department of Biotech-
nology and Genetic Engineering, Medical University of Silesia,
School of Pharmacy with the Division of Laboratory Medicine in Sosnowiec,
ul. JednoÅŻci 8, 41-200 Sosnowiec, Poland,
e-mail: ibednarek@sum.edu.pl

https://doi.org/10.1007/978-3-319-70063-2_2
14 Przemysáaw Jedrusik et al.
calculations have been mostly eliminated by including in the process a large com-
puting power of computer equipment. This paper refers to the problems connected
with behavior and relationships between lung cancer cells. Nowadays the problem
of cancer is very timely and important especially. Lung cancers are still at the fore-
front of medical conditions directly threatening the lives of people every year, in
2012 were over 410.000 new lung cancer cases diagnosed [5]. Understanding the
mechanisms of cell behavior significantly affects the effective methods of antitu-
moral immunotherapies. It was not possible to real analyze the different factors
that determine how migration and cell growth without the correct segmentation [8].
Analyzes of changes in saved medical images, are mainly based on the correct sep-
aration of visible cells from the background image. Separation of the course has
a lot of problems regarding the designation of individual cells as separate objects.
This article proposes the use of adaptive method of binarization including the re-
sults obtained after the binarization. The results allow properly direct algorithm by
eliminating objects, which cannot be the ones we are looking for, because of their
morphology.
2.2 Materials in experiment
The study of this work is based on microscopic analysis of medical images A549
lung cancer cells ATCC CCL-185TM. Lung cancer cell line originated from the
group ATCC, grown in accordance with the manufacturer’s instructions. Cell cul-
tures were routinely propagated in a humidified incubator Hera-Cell (Heraeus), at
37◦ C, in 5% CO2 atmosphere. Cells were grown in RPMI-1640 medium supple-
mented with 10% FBS and gentamicin (20 μg/ml ). Cells were seeded in 12-well
plates to reach confluency. Cells were allowed to growth and migrate for 24, 48
and 72 hours. Cells were visualized under inverted microscope (OLYMPUS BX-
60 microscope with DP50 digital camera) at indicated time points; magnification
100x. The acquisition of microscopic images was performed in the Department of
Biotechnology and Genetic Engineering Medical University of Silesia in Katowice.
Were registered 120 different images. All obtained images is provided in each of
the same area of the cell culture. The images in digital form are saved as a bitmap
with a resolution of 2080x1540 pixels. Pixel of acquired images was dimension of
approximately 10 nanometers in real.All images were analyzed using a proprietary
algorithm, written in Matlab version R2013b software with Image Processing Tool-
box version 8.3. The software was running on a computer with an Intel i5 processor
clocked at 2.6GHz, along with 8GB of DDR3 RAM, serial ATA hard disk drive and
operating system Microsoft Windows 10 Pro (version 1511, compilation 10586.63).
2 Automatic discover 15
2.3 Methods
Designed algorithm was divided into 2 stages. In a first step, images treated with the
methods of image analysis (pre-processing). A median filter was used to eliminate
artifacts in the input images. [1][2]. Filter size was matched to the size of objects
that were artifacts of a fixed local window size of 19x19 pixels [7]. The adaptive
binarization method was used in the next step with local window, which size was
adapted to the size of the areas, collections of pixels of brightness changes [3]. Typ-
ically, methods of binarization based on a global binarization threshold determined
for all pixels in the image. Modifications of these methods have been used in the
article, which analyzes an image with the division into regions, were used a local
adaptive thresholding techniques.
An example of input image can be found in Figure 1.
Fig. 2.1: Example source image of lung cancer cells.

0, if image(x, y) < th(x, y)
bw(x, y) = (2.1)
1, otherwise,
where th(x, y) was a threshold for each pixel at location (x, y).
In this paper was used the Sauvola’s method of binarization, based on the stan-
dard deviation and average value of the pixel at a specified point [9].
Fig. 2.2: Schematic of the proposed algorithm for automatic segmentation of lung
cancer cells.

s(x, y)
th(x, y) = m(x, y) 1 + k −1 , (2.2)
R
where threshold th(x, y) is computed using m(x, y) a mean, s(x, y) a standard devia-
tion of the pixel intensities in a w × w window and where R is the maximum value
of standard deviation and k is a positive parameter to modular according the ana-
lyzed image. In the next stage used morphological operations: erosion and dilation
with selected the structuring element, allowing remove objects that are considered
distortion.
In the next step specific cell parameters were determined, that represented their
individual properties. According to the proposed algorithm for each of the cells were
determined by the following parameters:
• surface area,
• the center of gravity,
• major and minor diagonals.
After considering these parameters were determined by the number of detected
cells and the distance between them, specifically between the centers of gravity of
the cells. The second stage of the algorithm was repeated until the appropriate values
determined parameters. The stop condition for the main loop has been reached a
local extreme value which depended on the average size of detected objects and
the maximum size of the object. Detected objects (cells) had a surface area less 2.5
times than the average cell size in the picture, and had size not less than half the
average size of all detected cells. Characteristic of biological cells is their growth
and division [6]. It was assumed that the cell before division for the next can be more
than 2 times larger than the average cell size, which at this stage does not exhibit
growth or division [4]. Specifically, the object was correctly detected as a single cell,
if the size was less than 2.5 times the average cell size. In analogy, the cells which
surface area were substantially smaller than the average value for all the cells, were
not taken into account. Such objects were treated as artifacts of the image.
∀c∈C (0.5 ·Cmeanarea < area(c) < 2.5 ·Cmeanarea ) , (2.3)

where C was set of all objects (cells), Cmeanarea was a mean area of all cells in set C.
The research were conducted on a group of 120 pictures cell cultures at differ-
ent times of proliferation. A wide group of input data allowed to obtain the results
shown in the next section of work, and to formulate conclusions.
2.4 Results
Implementation and execution of the designed algorithm for a example input images
allowed to obtain the data. Table 1 contains the partial data, collected for the sample
input image.
Analyzed table containing the results, it was found that the selected parameter
k is associated with an average size of objects and the size of the maximum de-
tected object in the image. The reduction of the average size of the cell area, also
decreases the surface area of the maximum field. Determined the appropriate value
of k, allowed to obtain an information regarding the change of the surface area of all
cells, the average distance between them, the average ratio of the diagonals, and the
amount of detected objects. The designated objects do not differ from each other in
terms of surface area. The shape of the objects was stretched slightly, the ratio of the
diagonals was less than 2. The distance between the cells was significantly higher
Table 2.1: Part of computed data, automated determine k - value
mean max cell mean major/minor

k cells
area [px] area [px] distance [px] axis
1 0,32 264 662 658 924 1,67
2 0,36 152 685 677 1026 1,88
3 0,22 225 1042 1038 699 1,86
4 0,24 164 1181 1172 887 1,97
5 0,18 282 860 854 1024 1,98
6 0,34 178 832 829 966 1,69
7 0,42 162 797 788 924 1,79
8 0,38 224 660 636 1029 1,66
9 0,26 197 786 781 1104 1,79
than the average cell size, which leads to the conclusion that there were blank areas
between cells.
Fig. 2.3: a) mean cells area and area of largest detected cell, b) the absolute value of
the difference between mean cells area and largest detected cell
The figure 3 shows example of the dependence of the average cell size of the area
and the area of the largest cell in the subsequent steps of the loop algorithm.
The figure 4 shows example of the mean distance between cells, mean major
and minor axis ratio and number of cells on iteration steps. For the purposes of that
study, loop was stopped at 20-th iteration.
The figure 5 shows the optimum automatic detection of objects - cell lung cancer.
Fig. 2.4: a) mean distance between cells, b) mean major and minor axis ratio, c)
number of detected cells
Fig. 2.5: Example of automatic segmentation.
2.5 Discussion
Execution of the algorithm loop allowed to automatically determine the values of k

occurring in the formula (2). In the next iterations, the average value of the cell area
identified as targets has been relevant to the area of the largest object detected. This
means that in each successive iterations of the objects that were detected as signif-
icantly larger than the other were the appropriate parameters change binarization
detected as a greater number of smaller objects. End of the loop gave eliminating
all objects significantly larger than the mean. New objects do not differ in size from
mid-sized cells. At the same time the objects significantly smaller than the average
size of the cell surface have been removed from the set of designated objects. In ad-
dition, determined the objects characterized by reduced the diagonals ratio, where
it was less than 2. This means that the objects are not too elongated. It was also
observed the relationship between the detection of a valid objects, and the average
distance between them. Changing the average distance between the cells changed in
a linear with minimal inclination of the trend line to the x-axis.
2.6 Conclusions
The proposed algorithm allows automatic adjustment of imaging parameters of cell

cultures. The next steps of the algorithm was based on the information obtained
in the previous steps. Parameters determining the results obtained were: the aver-
age area of cells, the surface area of the cell with the largest distance between the
centers of gravity of the cells, the ratio of the diagonals of the set of cells and the
number of assigned cells. The parameters clearly define the appearance of objects
obtained, the distribution of the image. The entire process runs automatically, al-
lows the analysis of a large number of images. As a tool to assist assessment of
the behavior of tumor cells can be widely applied not only to lung cancer cells, but
also to any other cell line. Naturally you must specify the biological properties of
growing and migrating cells. Determination of the optimal number of cells further,
while maintaining the condition that the cells can overlap or merge. They may not
be properly recognized. Creation of this tool also allows to specify automatically
confluent cells value, which is a very important determinant performed by the in-
vestigator. Information about the confluence was included in the number of cells,
but especially in the distance between them. This means that determination of the
size of the free space between cells - as one of the parameters of this study. This of
course will require the implementation of further research.
References
1. Acharya, T., Ray, A.K.: Image Processing - Principles and Applications. Wiley InterScience;
2006
2. American Cancer Society. Global Cancer Facts & Figures 2nd Edition. Atlanta: American Can-
cer Society; 2011.
3. Angenent,S., Pichon, E., Tannenbaum, A.: Mathematical methods in medical image processing.
Bulletin of the American mathematical society, 43, 365-396; 2006
4. Elizabeth Flate, John R. D. Stalvey: Motility of select ovarian cancer cell lines: Effect of
extracellular matrix proteins and the involvement of PAK2. Int J Oncol. 2014 Oct; 45(4):
1401âĂŞ1411.
5. J. Ferlay, E. Steliarova-Foucher, J. Lortet-Tieulent, S. Rosso, J.W.W. Coebergh, H. Comber, D.

Forman, F. Bray; Cancer incidence and mortality patterns in Europe: Estimates for 40 countries
in 2012; European Journal of Cancer (2013) 49, 1374âĂŞ 1403
6. Jianping Peng, Ganesan Ramesh, Lin Sun, and Zheng Dong: Impaired Wound Healing in Hy-
poxic Renal Tubular Cells: Roles of Hypoxia-Inducible Factor-1 and Glycogen Synthase Kinase
3β /β -Catenin Signaling. J Pharmacol Exp Ther. 2012 Jan; 340(1): 176âĂŞ184.
7. Koprowski, R., Korzyńska, A., Zieleźnik, W., Wróbel, Z., Małyszek, J., St˛epień, B.,Wójcik,
W.: Influence of the measurement method of features in ultrasound images of the thyroid in the
diagnosis of Hashimoto’s disease. BioMedical Engineering OnLine, 11:91 (2012)
8. Nilendu C Purandare, Venkatesh Rangarajan; Imaging of lung cancer: Implications on staging
and management; Indian J Radiol Imaging. 2015 Apr-Jun; 25(2): 109âĂŞ120
9. Sauvola J., Pietikainen M.: Adaptive document binarization; Pattern Recognition 33(2), 2000,
p. 225-236
Chapter 3
Breast Cancer Segmentation Method
in Ultrasound Images
Marta Galińska1 , Weronika Ogiegło2 , Agata Wijata2 ,

Jan Juszczyk2 , and Joanna Czajkowska2
Abstract.
The most common type of cancer among women is breast cancer. The early di-
agnosis is crucial in a treatment process. The radiology support system in the diag-
nostic process allows faster and more accurate radiographic contouring. The aim of
the paper is to present a new method for ultrasound image segmentation of breast
lesions. The segmentation technique is based on active contour models whereas
anisotropic diffusion is used for preprocessing. The Dice Index calculated in most
of analyzed cases was greater than 80%. Delineation of the tumor can also be used to
calculate the size and volume automatically, and shortened the time of the diagnosis.
Keywords: US image segmentation, active contour, breast cancer
3.1 Introduction
Breast cancer is the second most common cancer nowadays and first among women,
10% of all new cases are breast cancers [1].
Currently, breast tumor prophylaxis mainly consists of screening mammography.
It is not a procedure that is indifferent to the patient’s health because of exposing
them to radiation. Moreover, subjective radiological diagnosis omits up to 10% le-
sions, especially when breast tissue is dense [2]. According to Jackson (1993), there
are also many false positive results of the mammography confirmed by a biopsy [3].
The second method of detecting lesions is ultrasonography (US). It is widely
available and non-invasive procedure. The huge advantage of the method is that it is
independent of breast density. However, US images are difficult to interpret due to
Silesian University of Technology, Faculty of Automatic Control, Electronics and Computer Scie-
nce, Gliwice, Poland,
· Silesian University of Technology, Faculty of Biomedical Engineering, Zabrze, Poland
e-mail: marta.galinska@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_3
24 Weronika Ogiegáo et al.
the uniqueness of the probe location during the procedure. An additional difficulty
is a presence of the noise and acoustic shadows in the images.
In 2004 Yu-Len Huang and Dar-Ren Chen have published a study on 60 patients
(included 39 carcinomas and 21 benign breast tumors) [4]. The system used textural
analysis to determine inputs of the neural network. Based on an autocovariance
coefficients, the tumors have been classified by self-organizing maps. After that,
the watershed algorithm has been used to the lesion segmentation. Due to the high
efficiency of the system, it has been applied in hospitals.
In 2008 N. Ravia Shabnam Parveen presented the new method of tumor segmen-
tation in US images [5]. The author used simple filters and morphological operations
like Gaussian filter, thresholding, holes filling, opening and reconstruction to deter-
mine the region of interest (ROI). Based on the shape and boundaries of the ROI a
classification has been performed.
The other study in US image classification has been accomplished using 87 im-
ages. The applied approach consisted of four steps: preprocessing (contrast enhance-
ment, denoising), segmentation (Markov random fields), feature extraction and se-
lection (textural, fractal and histogram analysis), classification and evaluation (using
fuzzy support vectors machine) [6].
The approach of genetic algorithms (GA) to US image segmentation is also found
in the literature. Gomez [7] extracted 22 morphological features using GA. After
histogram equalization and anisotropic diffusion, the lesion has been segmented by
the watershed algorithm. Artificial neural network has been used for classification
and feature reduction. Finally, 7 of 22 features has been selected. The study has been
performed on 641 US images.
Sasikala et al. in 2014 have published a paper contained results of their work on
breast tumors classification [8]. The basis of the method was the simultaneous lesion
segmentation in US and elastographic images. Area of the tumor has been separated
using the level set segmentation technique. Speckle Reducing Anisotropic Diffusion
(SRAD) has been used for preprocessing. A feature comparison between three types
of lesions has been performed.
All of the foregoing articles aimed at classification if breast masses are malignant
or benign. Automated or semi-automated segmentation was only for determining the
region of the lesion. Another approach was proposed by Chang at al. [9]. The au-
thors attempted to create the 3D model of the lesion. Preprocessing step contained
watershed algorithm, thresholding and morphological operations. They used 3D ac-
tive contours for fit improvement of the initial model. The segmentation results were
rewarding and its visualization was useful for radiologists.
In 2014 Lo proposed an Automated whole Breast US (ABUS) imaging systems
[10] and the computer aided detection system of ABUS data. The image process-
ing contained contrast enhancement, denoising, watershed algorithm and textural,
intensity and morphological feature analysis. During clinical tests, 100% sensitivity
has been obtained.
The purpose of this paper is to propose a new method for semi-automated seg-
mentation of the breast lesions in US images. It has a significant influence on short-
3 Breast Cancer Segmentation 25
ening of breast diagnostic time. The automation of the segmentation process con-
tributes to more objective, independent on radiologist and repeatable diagnosis.
After a brief introduction materials and methodology are presented in Section 2.
Section 3 contains results and their discussion, and a short conclusion is presented
at the end of the paper.
3.2 Materials and Methods
The analysed dataset consists of two subsets. The first one contains 6 synthetic ul-
trasound images of a phantom recorded by using Philips iU22 system with L12-5
linear probe. The phantom consists of hydrogel sphere inserted into the animal mus-
cle tissue. The hydrogel echogenicity is similar to the echogenicity of breast cyst.
The muscle was used as external tissue surrounding the lesion.
The second subset contains 6 clinical images of breast cancer. The data were
recorded by using Hitachi EZU-MT25-S1 system with linear probe L65. The images
were acquired during a routine US examination performed by a radiologist. The
penetration depth of the ultrasound beam was in the range of 30 to 50 mm.
Figure 3.1 presents workflow of the segmentation. The first branch represents
the loading of image data and anisotropic diffusion [11] which is used as a filter.
The second branch consists of marking startpoints and interpolation between them.
Process of semi-automated segmentation starts from at least 4 seedpoints on bor-
der between lesion and background. The interpolation step is used to automatically
determine the position of the other startpoints located near the border. First, ap-
proximation of lesion contour is performed. Next, an active contour method [13] is
applied to the filtered image. The contour obtained by the method yields the lesion
mask.
Fig. 3.1: Block diagram of the proposed method with an illustration of all steps.
The employed anisotropic diffusion is an iterative algorithm widely used for

noise reduction. This technique is edge-preserving which yields sharp boundaries
and smooth homogeneous areas [12]. It involves solving partial differential equa-
tions and is described by
I t+1 (x) = I t (x) + λ ∑ (ct (x, y) · (I t − I t (y))) (3.1)

y∈N(x)
where I is the intensity of the image, t is the number of iteration, x is the analyzed
pixel, λ is the diffusion rate, N(x) is neighborhood of the pixel x and ct (x, y) =
g(I) is diffusion coefficient.
The diffusion coefficient c consists of two components. First, c1 gives priority to
the edges (eq. 3.2) and second, c2 prioritize the intensity (eq. 3.3),

| I|2
c1 (x, y) = exp (3.2)
2k2
1
c2 (x, y) = (3.3)
1 + |I|
2
2k2
where k is the edge magnitude parameter and is a local gradient.
The next step is the active contouring [13]. The active contour model is an
energy-minimizing spline which is fitting to the edges of the structure. Nodes lo-
cation are modified by internal and external forces. The process of moving nodes
is iteratively repeated until the termination condition is met. For each point v(s) =
(x(s), y(s)) a snake energy Esnake functional is given as
1
Esnake = Eint (v(s)) + Eimage (v(s)) + Econ (v(s))ds, (3.4)
0
where Eint , Eimage and Econ are internal, image and external constraint forces, re-
spectively.
The internal spline energy Eint is calculated as
1 1
Eint = (α (s)|vs (s)|2 ) + (β (s)|vs (s)|2 ), (3.5)
2 2
where α is a stretch sensitivity parameter and β is sensitivity to the amount of
curvature in the snake. The image energy Eimage is defined as
Eimage = wline Eline + wedge Eedge + wterm Eterm , (3.6)

where wline , wedge and wterm are weights, Eline = I(x, y), Eedge = −| I(x, y)|2
and
Eterm depends on gradient direction.
The last step is labeling procedure. Due to the fact, that the applied active contour
is a closed curve, the pixels inside it can be described as belonging to the mask. The
remaining pixels are the background.
3.3 Results and Discussion
To verify the effectiveness of the developed method, two following measures are
calculated:
• Dice index as a measure of the similarity of two binary masks (obtained by the
method and by the expert delineation) [14],
• Hausdorff distance as a maximum distance of a mask to the nearest point be-
longing to the gold standard [15].
Fig. 3.2: Dice Index and Hausdorff Distance for each analyzed image.
Figure 3.2 shows differences between accuracy of the segmentation process per-
formed on synthetic and clinical images. The baseline indicates the median of the
entire set. The median values of Dice Index were 0.85 and 0.8 for synthetic and
clinical data, respectively. It shows that the method works well with both synthetic
and clinical images. Up to 2 out of the clinical cases have achieved better results
than synthetic ones. The real lesions have various size and shapes (Fig.3.4) in op-
position to synthetic US images of the hydrogel sphere which shape is regular and
round (Fig.3.3). The noise and artifacts caused distortion and irregularity of borders.
Differences between results obtained from clinical and synthetic data may arise be-
cause of lower echogenicity and higher homogeneity of the hydrogel compared to
clinical lesions. These differences directly affect the efciency of the algorithm.
In Table 3.1 numerical results for clinical and synthetic data are presented. Table
3.2 contains a comparison of the Dice Index statistics for both groups. Mean and
median values for the synthetic images were higher than values for the clinical cases.
Clinical images were less homogeneity than the synthetic dataset. A comparison
between Hausdorff Distances is shown in Table 3.3. Results are shown in pixels, due
to the characteristics of the method. Depending on the image resolution, different
segmentation accuracy was obtained. The results were similar to the analysis of the
Dice Index. In the clinical data, mean and median values of the Hausdorff Distance
Table 3.1: Numerical results for all cases.
Synthetic case 1 2 3 4 5 6
Dice Index 0.92 0.88 0.78 0.79 0.84 0.90
Hausdorff Distance 4.89 5.83 6.26 5.92 5.61 4.98
Clinical case 7 8 9 10 11 12
Dice Index 0.87 0.80 0.79 0.82 0.64 0.87
Hausdorff Distance 8.77 9.43 6.86 10.44 14.18 5.74
Table 3.2: Dice Index statistics.
Dice Index Synthetic data Clinical data

Mean 0.85 0.80
Median 0.86 0.81
Std 0.058 0.085
indicate slightly less fit of the obtained mask. The standard deviation of the clinical
results was almost 6 times higher than in synthetic cases.
Table 3.3: Hausdorff Distance statistics.
Hausdorff Distance Synthetic data Clinical data

Mean 5.58 9.24
Median 5.72 9.10
Std 0.54 2.97
The Hausdorff Distance increases with the degree of irregularity of the edges.
Because of that, almost all of the synthetic images have lower values than clini-
cal images. Dice Index in most cases was greater than 80%. Exemplary results are
presented in Fig. 3.3 and 3.4. Image numbers correspond to the numbers in Table
3.1. The remaining results are similar to those presented. Only in one of the clinical
cases, it was 64%. It may be caused by the large size and specific shape of the lesion
and very torn edges. The original image, expert delineation and obtained result for
this case are shown in Fig. 3.5.
3.4 Conclusion
The main purpose of the study was to determine borders of the lesion in ultrasound
images of breast tissue. The obtained results indicate that lesions in most cases have
been delineated properly by the proposed method. For 11 of the 12 cases analyzed,
the method correctly identified the edges of the lesion. Usage of the proposed al-
gorithm might have a positive impact on time of the diagnosis. Please note that the
(a) (b) (c)
Fig. 3.3: Sample results of segmentation performed on synthetic images (cases 1, 2,

4).
(a) (b) (c)
Fig. 3.4: Sample results of segmentation performed on clinical images (cases 7, 9,

12).
(a) (b) (c)
Fig. 3.5: The worst case: 5(a) original image, 5(b) obtainted mask and 5(c) expert
delineation.
nal diagnosis depends on the radiologist and semi-automated method is more ac-
ceptable by him. An experience of an expert is impossible to be replaced by the
computer. However, we can try to help radiologist by determining the approximate
boundaries automatically. If the boundaries are not correct, a radiologist can observe
it easily and perform the delineation by himself. Determination of the tumour con-
tour can also be used to automatically calculate the size and volume. The developing
of tumor segmentation method will be continue further and it will be tested on more
cases.
Acknowledgements
This research was supported by the Polish Ministry of Science and Silesian Uni-
versity of Technology statutory nancial support for young researchers BKM-
510/RAu-3/2017. The funders had no role in study design, data collection and anal-
ysis, decision to publish, or preparation of the abstract.
References
1. Ferlay, J., Hery, C., Autier, P., Sankaranarayanan, R.: Global burden of breast cancer. Breast
cancer epidemiology. Springer New York, 1–19 (2010)
2. Dennis, M., Parker, S., Klaus, A., Stavros, A., Kaske, T., Clark, S.: Breast Biopsy Avoidance:
The Value of Normal Mammograms and Normal Sonograms in the Setting of a Palpable Lump
1. Radiology 219, 1, 186–191 (2001)
3. Jackson, V., Hendrick, R., Feig, S., Kopans, D.: Imaging of the radiographically dense breast.
Radiology 188(2), 297–301 (1993)
4. Huang, Y., Chen, D.: Watershed segmentation for breast tumor in 2-D sonography. Ultrasound
in Medicine & Biology 30(5), 625–632 (2004)
5. Parveen, N.: Segmenting tumors in ultrasound images. International Conference on Computing,
Communication and Networking, St. Thomas, VI, 1–5 (2008)
6. Shi, X., Cheng, H.D., Hu, L., Ju, W., Tian, J.: Detection and classication of masses in breast
ultrasound images. Digital Signal Processing 20(3), 824–836 (2010)
7. Gomez, W., Rodriguez, A., Pereira, W., Infantosi, A.: Feature selection and classier perfor-
mance in computer-aided diagnosis for breast ultrasound. Emerging Technologies for a Smarter
World (CEWIT), 10th International Conference and Expo on. IEEE (2013)
8. Sasikala, S., Kirthika, B., Malathi, P.: Feature Extraction and Analysis of Breast Lesion in
Ultrasound B Mode and Elastography. International Journal of Advanced Research in Computer
Science and Software Engineering 4(1), 355–359 (2014)
9. Chang, R., Wu, W., Moon, W., Chen, W., Lee, W., Chen, D.R.: Segmentation of breast tumor
in three-dimensional ultrasound images using three-dimensional discrete active contour model.
Ultrasound in Medicine &Biology 29(11), 1571–1581 (2003)
10. Lo, C., Chen, R., Chang, Y., Yang, Y., Hung, M., Huang, C., Chang, R.: Multi-dimensional tu-
mor detection in automated whole breast ultrasound using topographic watershed. IEEE Trans-
actions on Medical Imaging 33(7), 1503–1511 (2014)
11. Yu, Y., Acton, S.: Speckle reducing anisotropic diffusion. IEEE Transactions on Image Pro-
cessing 11(11), 1260–1270 (2002)
12. Perona, P., Malik, J.: Scale-space and edge detection using anisotropic diffusion. IEEE Trans-
actions on Pattern Analysis and Mmachine Intelligence 12(7), 629-639 (1990)
13. Kass, M., Witkin, A., Terzopoulos, D.: Snakes: Active contour models. International Journal
of Computer Vision 1(4), 321–331 (1988)
14. Seghier, M., Ramlackhansingh, A., Crinion, J., Leff, A., Price, C.: Lesion identication us-
ing unied segmentation-normalisation models and fuzzy clustering. Neuroimage 41(4), 1253–
1266 (2008)
15. Huttenlocher, D., Klanderman, G., Rucklidge, W.: Comparing images using the Hausdorff dis-
tance. IEEE Transactions on Pattern Analysis and Machine Intelligence 15(9), 850–863 (1993)
Chapter 4
Detection and Tracking of the Biopsy Needle
Using Ultrasound Images
Agata Wijata1 , Żaneta Ranosz2 , Marta Galińska2 ,

Jan Juszczyk2 , Joanna Czajkowska2
Abstract. The aim of this work is to develop a method of detecting and tracking
the needle tip using only two-dimensional ultrasound images. A novel method based
on Hough transform, Shock filter and Gabor filter is proposed. The algorithm em-
ploys a US image to extract the needle tip. First derivative analysis is used for verifi-
cation and correction the tip coordinates. The proposed method including the needle
tip detection correction has an efficiency of 80%. The method may support radiolo-
gist during a core needle biopsy.
Keywords: core needle biopsy, needle tip detection, ultrasound image
4.1 Introduction
Biopsy procedure is one of diagnostic techniques for cancer patients. The procedure
is performed to confirm or exclude the pathology based on the microscopic assess-
ment of the collected tissue fragment. There are various biopsy methods which are
used in dependence on the type of material being taken, test conditions, lesion lo-
cation and patient condition. The most frequently performed procedures include:
fine needle aspiration biopsy, core needle biopsy, surgical biopsy and intraoperative
biopsy [2]. Fine needle and core needle biopsy are performed under ultrasound con-
trol. During these procedures, 3-5 tissue fragments comming from different parts of
the tumor are taken to determine the heterogeneity of the lesion [1].
The fine needle biopsy involves puncturing the tumor using a 0.5-0.9 mm di-
ameter needle and collecting the cytological material for microscopic evaluation.
This biopsy is used to examine tumor lesions detected by palpation, radiology, ul-
Silesian University of Technology, Faculty of Automatic Control, Electronics and Computer Scie-
nce, Gliwice, Poland
e-mail: agata.wijata@polsl.pl · Silesian University of Technology, Faculty of Biomedical
Engineering, Department of Informatics and Medical Equipment, Gliwice, Poland

https://doi.org/10.1007/978-3-319-70063-2_4
34 Agata Wijata et al.
trasonography and scintigraphy, irrespectively of the location of the tumor in the pa-
tient body. It is particularly useful in the diagnosis of breast, prostate, bone marrow
and lymph node metastases. It is also employed in the diagnosis of tumors of the
salivary glands, liver, thyroid, and lung [3]. Compatibility needle aspiration biopsy
with a final diagnosis oscillates between 80% and 95% for perceptible breast, lung,
thyroid or bone tumors. For non-perceptible tumors, this is between 60% and 70%
in case of tumors of lymph nodes and pancreas and 80%-90% for liver [2].
A small volume of material collected during the biopsy is sometimes insufficient
in the diagnostic process. Therefore, core needle biopsy is also commonly used.
This biopsy procedure requires the use of a needle gun with a needle diameter of
1.2-2.3 mm. A radiologist collects 3-5 samples under ultrasound control. Material
samples are subjected to histopathological assessment. Core biopsy is performed in
case of palpable tumor of the breast, bone, prostate, liver, lymph nodes, lung and soft
tissue. It is also used during surgery to diagnose pancreatic tumors. The diagnostic
sensitivity of core needle biopsy is 80%-90%. The important diagnostic informa-
tion of this method is a confirmation of the pre-diagnosed leptospirosis. A biopsy
indicating mild lesion diagnosis requires further investigation [1, 4].
The biopsy goal is to take the tissue material from the assessed lesions. The dis-
advantage of the procedure is the lack of control over the injection site and the way
the material is collected. Incorrect insertion of a needle into patient body creates the
risk of picking up the wrong tissue or damaging the internal organs. A radiologist
using a two-dimensional ultrasound image can observe only a cross section of the
lesion during the biopsy. The difficulty is also the simultaneous observation of the
lesion and the needle in the 2D image and determining the point of collection of
the material for the analysis. The use of the tracking system and 3D breast model
allows determining biopsy sites and visualization of the procedure, which increases
the effectiveness of the biopsy [5]. However, the marker placed on the biopsy gun
does not permit a clear indication of the position of the needle tip. The reason for the
inaccuracy is a significant flexibility of the biopsy needle. Therefore, it is important
to track the biopsy needle in ultrasound (US) images.
In the US scans, the biopsy needle appears as a continuous straight line, usually
brighter than the surrounding tissue. Therefore, it can be recognized using Hough
transform [6]. It requires finding the dominant line in the image and indicating the
end point of the found straight (needle tip). The Hough transform is a relatively sim-
ple and effective technique in case of good quality images with well visible needle.
Unfortunately, if the needle is not well visible indicating its position is impossible.
The needle detection is performed on the basis of deformation of tissues which
are located around the biopsy needle track. The occurrence of local tissue defor-
mation is determined by the distribution of local gradient or edge orientation. The
extraction of these features is based on Histogram of oriented Gradients (HoG). The
designated tissue features are classified using Weighted Fuzzy C-Means algorithm.
This approach allows segmentation of the area deformed by the inserted needle [7].
A combination of the HoG features with Kernelized Weighted C-Means clustering
presented in [8] also allows the detection of tissue deformation area. It indicates the
4 Detection and Tracking of the Biopsy Needle 35
needle position in 2D US images in case of hardly visible needle. Unfortunately, it

does not indicate the needle tip.
In [9], needle detection is performed using a Gabor filter and morphological op-
erations. After the estimation of angle of needle insertion, the image is filtered using
the Gabor filter. The filter enhances linear elements in a specific direction and al-
lows to indicate the needle tract and its tip. The method is characterized by high
efficiency but requires the visibility of the needle on images.
The purpose of this work is to develop a method of detecting and tracking the nee-
dle tip using two-dimensional ultrasound images. A proposed method is based on
combination of Hough transform, Shock filter and Gabor filter. The needle tip co-
ordinates are verified and corrected using first derivative analysis. The algorithm
employed to the US image extracts the tip of the needle.
The paper is organized as follows. In section 2 the applied methods are described.
Section 3 presents results and discussion. Finally, conclusions and plans for further
work are presented in section 4.
4.2.1 Materials
The data were acquired using Philips iU22 Ultrasound Machine with linear trans-
ducer Philips L12-5. The experiments were performed on phantoms made of pork
shoulder. In total 8 sets of images were acquired. The detection of needle tip was
performed in MATLAB environment.
4.2.2 Methods
The proposed method detects the needle tip on the series of two-dimensional US
image (Fig. 2(a)).
4.2.2.1 Dominant line detection
In literature, dominant line detection on images is often performed using Hough

transform [6]. In our approach, the Hough transform is used to locate the area with
visible biopsy needle. The ultrasound image is represented by pixels with Cartesian
coordinates (x, y). As reported in [10], each line can be represented using Hesse
normal form
ρ = x cos θ + y sin θ (4.1)

Fig. 4.1: Block diagram of the algorithm
where ρ is the distance from the origin to the closest point in this line, and θ is the
angle between the x axis and the line which connects the origin with this point.
Point coordinates are applied to the parameter space (θ , ρ ). The family of curves
is determined using a sinusoid at each point. If two curves intersect at two different
points in the parameter space, it means that analyzed points lie on a straight line
in Cartesian coordinates. The line in the image is detected after exceeding defined
number of intersections. The dominant line found in the image represents the needle
and the needle tip is searched along it. Based on information about the trajectory
of the needle movement and the angle of insertion of the needle in relation to the
horizontal direction, it is possible to determine the region of interest (ROI) and angle
of insertion for further analysis.
4.2.2.2 Edge sharpening
For edge sharpening, the Shock filter [11] is employed. This filter uses a local di-
latation or erosion process to modify pixels based on belonging to the maximum or
minimum impact zone:
t−1
Δ f (xi , y j ) < 0 =⇒ f t (xi , y j ) = I t−1 ( f (xi , y j )) ⊕ D
(4.2)
Δ f t−1 (xi , y j ) > 0 =⇒ f t (xi , y j ) = I t−1 ( f (xi , y j )) D
for each pixel f t (xi , yi ) of a two-dimensional image { f t (xi , y j }i=1,M; j=1,N at itera-
tion t, where Δ f t (xi , yi ) is Laplacian computed at pixel (xi , yi ). D is a disk-shaped
structuring element of radius 1. Symbols ⊕ and correspond to dilation and ero-
sion, respectively. This filter is often used to remove noise and sharpen the edges on
ultrasound, x-ray, computed tomography and magnetic resonance images [11, 12].
As a result, after the filtration the needle image is sharpened and a large amount of
noise is suppressed.
In the next step, the ROI is applied to the image filtered using Shock filter. The
result of the operation is analyzed using a Gabor filter.
4.2.2.3 Needle enhancement
The needle is enhanced using Gabor filter, which is employed to reinforce structures
of a particular direction [9]. A two-dimensional Gabor filter is defined by:
x 2 + γ 2 y 2 x
g(x, y; λ , θ , ψ , σ , γ ) = exp(− ) cos(2π + ψ) (4.3)
2σ 2 λ
where x = x cos θ + y sin θ ; y = −x sin θ + y cos θ ; λ and θ are the wavelength
and orientation of the sinusoidal plane wave, respectively; ψ is the phase offset;
σ is the standard deviation of the Gaussian envelope along the x- and y-directions,
respectively; γ is the spatial aspect ratio [13]. The result of filtering is suppres-
sion of the noise, whose orientation angles are different from the angle of insertion
of the needle. The needle is clearly visible in the image.
4.2.2.4 Needle tip coordinates determination
The thresholding is performed in order to maintain the most enhanced filtration ele-
ments. Then, the biggest element which represents the needle is selected. The needle
center is calculated. Based on it, the needle is reconstructed. The tip is the extreme
point of the segmented needle. The coordinates of the extreme point are taken as
the coordinates of the needle tip. The result of needle tip coordinates determination
is presented in Figure 2(b).
4.2.2.5 Correcting process
The last step of the methodology is an attempt to detect the wrong results and au-
tomatically improve them. For this purpose, the graph of changes in coordinates
(x, y) of potential needle tip point is analyzed. Invalid results reported for the point
are characterized by a sudden decrease or increase in distance value (Fig. 4.3).
These locations are found using first derivative analysis. Detection of violent po-
(a) (b)
Fig. 4.2: Exemplary, 2(a) 2D US image and 2(b) result of the detection of needle tip
sition changes is done by thresholding (t = 10). The new value is determined by

linear interpolation between preceding and following correct points.
Fig. 4.3: Comparison of X coordinate change for an example series

Verification of the algorithm was performed on 8 series of core needle biopsy im-
ages. For recorded data, the expert manually marked the correct position of the nee-
dle tip. Expert evaluation was used to assess the effectiveness of the algorithm.
The comparison was made using images with visible needle. Due to high level of
the noise, the first 10 images of each series were not included in the assessment. The
results were compared by calculing the Euclidean distance between the point found
automatically and indicated by the expert. The determined distances for each im-
age were compared to the 12-pixel threshold, which was empirically chosen as the
double needle thickness. The results of comparison are presented in the first row
of Table 33.1. The average matching of the proposed method to the expert rating is
48.60%.
Table 4.1: Impact of first derivative analysis
series of images 1 2 3 4 5 6 7 8
% of fit 76.4 61.69 56.27 52.98 49.44 32.59 30.67 28.77
% of fit using f 80.34 62.21 62.52 53.89 49.44 36.62 31.33 29.23
The effect of using first derivative analysis is to improve efficiency of the pro-
posed method (second row of Tab. 33.1). The average matching of the improved
method to the expert rating is 50.70%. The presented comparison of algorithm re-
sults with expert judgment is characterized by high variability. An analysis of the dis-
tance between the obtained result and the point indicated by the expert was per-
formed to assess the effectiveness. The distance between a point determined by the
proposed method and the expert point was measured. The operation was repeated
for each image. The distance was calculated in millimeters, according to the known
needle diameter (2.2 mm) and the measured mean diameter of the needle in the
images (6 pixels). Average distance and standard deviation were calculated for all
series analyzed. The results are summarized in Table 26.3.
Table 4.2: Average values and standard deviation distance between the results
of the algorithm and the expert’s assessment in pixels
series of images 1 2 3 4 5 6 7 8
average 4.2 14.8 8.0 14.6 8.3 19.9 49.7 21.3
std 6.9 20.5 8.5 23.3 9.9 18.4 43.0 17.0
The analysis of results of the comparison shows that the value of the average
distance and standard deviation are characterized by low values for the best series.
On this basis, it is possible to conclude that there is agreement between the results
obtained by the automatic method and the expert outlines. For low performance
results, the average distance and standard deviation are high. It shows that the algo-
rithm returns extremely good or extremely bad values.
The results of the algorithm were promising in 5 of 8 cases (50% ± compliance
with an expert is considered a success). It means that proposed method can support
radiologist during biopsy procedure.
4.4 Conclusion
The method of detection and tracking of the needle presented in the work accom-
plishes the task in several stages. The first is the use of the Hough transform to deter-
mine the needle trajectory in a series of images and the angle of insertion of a needle.
Based on the trajectory, the region of interest is designated. Each image of the series
is subjected to a Shock filter and a Gabor filter to emphasize the needle on the ul-
trasound images and eliminate the noise. Binary and morphological operations are
used to obtain a needle binary mask. In the last stage the needle tip is determined
by the mask. In the best case the result was 76% for the comparing results of the
algorithm and the expert outlines. The evaluation is difficult because of invisibility
of the entire needle on each of the analyzed images. Based on the analysis of the
first derivative, abnormal needle detection moments were corrected. Again, suma-
rized result was compared with the expert analysis and the result was 80% of effec-
tiveness. An important stage of work with the use of ultrasound images is filtering
and noise reduction because the quality of input data determines the effectiveness
of the segmentation. The quality of the apparatus used during a biopsy is also an
important factor.
Developed method is the preliminary study on the effective method of detecting
needles on images in real time. This can significantly affect the convenience of the
procedure and its effectiveness. The result may be a higher percentage of correctly
diagnosed lesions and more effective oncology treatment.
Acknowledgements
This research was supported by the Polish Ministry of Science and Silesian Uni-
versity of Technology statutory financial support for young researchers BKM-
510/RAu-3/2017. The funders had no role in study design, data collection and anal-
ysis, decision to publish, or preparation of the abstract.
References
1. Ślubowski, T., Ślubowska, M., Wojciechowski, A.: Techniki Diagnostyczne w Raku Piersi.
Cześć II: Metody fizyczne i biopsyjne. Ginekologia Polska, 479–483 (2007)
2. Skokowski, J.: Encyklopedia Badań Medycznych. Badanie Histopatologiczne. Wydawnictwo

Medyczne MAKmed, Gdańsk (1996)
3. Szopiński, K., Szopińska, M.: Current Opinion on Imaging and Biopsy of Breast Masses. Acta
Clinica 2(3), 273–294 (2002)
4. Rocha, R., Pinto, R., Tavares, D., Goncalves, C.: Step-by-Step of Ultrasound-Guided Core-
Needle Biopsy of the Breast: Review and Technique. Radiologia Brasileira 46, 234–241 (2013)
5. Treepong, B., Tanaiutchawoot, N., Wiratkapun, C., Suthakorn, J.: On the Design and Develop-
ment of a Breast Biopsy Navigation System: Path Generation Algorithm and System with Its
GUI Evaluation, Biomedical and Health Informatics, 273–276 (2014)
6. Hong, J., Dohi, T., Hashizume, M., Konishi, K., and Hata, N. An Ultrasound-Driven Needle-
Insertion Robot for Percutaneous Cholecystostomy. Physics in Medicine and Biology, 441–455
(2004)
7. Czajkowska, J., Juszczyk, J., Pycinski, B., and Pietka, E.: Biopsy Needle and Tissue Deforma-
tions Detection in Elastography Supported Ultrasound. Information Technologies in Medicine,
85–96 (2016)
8. Czajkowska, J., Pycinski, B., and Pietka, E. HoG Feature Based Detection of Tissue Deforma-
tions in Ultrasound Data. Engineering in Medicine and Biology Society (EMBC), 2015 37th
Annual International Conference of the IEEE, 6326–6329 (2015)
9. Kaya, M., and Bebek, O.: Needle Localization Using Gabor Filtering in 2D Ultrasound Images.
IEEE International Conference on Robotics & Automation (ICRA), 4881–4886 (2014)
10. Duda, R., and Hart P.: Use of the Hough transformation to detect lines and curves in pictures.
Artificial Intelligence Center. SRI International 15(1), 11–15 (1971)
11. Vacavant, A., Ali, A., Grand-Brochier, M., Albouy-Kissi, A., Alfidja, A., Chabrot, P., and
Boire, J.: Smoothed Shock Filtered Defuzzification with Zernike Moments for Liver Tumor
Extraction in MR Images. IEEE Image Procesing Theory, Tools and Aplications, 17–21 (2015)
12. Vacavant, A., Albouy-Kissi, A., Menguy, P. Y., and Solomon J.: Fast Smoothed Shock Fil-
tering. Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012),
182–185 (2012)
13. Rangayyan, R., and Ayres, J: Gabor Filters and Phase Portraits for the Detection of Architec-
tural Distortion in Mammograms. Medical & Biological Engineering & Computing 44, 883–
894 (2006)
Chapter 5
Detection of Wet Age-related Macular
Degeneration in OCT Images: A Case Study
Anam Haq and Szymon Wilk
Abstract. Progress in medical imaging and computer vision has enabled us to rely
on machines for the detection or diagnosis of many diseases, including eye-related
problems. One of them is wet age-related macular degeneration (wet AMD) which is
a type of age-related macular degeneration. Wet AMD causes the detachment of reti-
nal pigment epithelium layer (RPE) – a condition referred to as pigment epithelium
detachment (PED) – and also creates fluid fill region called choroidal neovascular-
ization (CNV). In this paper we present a case study of detecting wet ADM in OCT
images. We used a set of 51 images – 21 of sick eyes and 30 of healthy eyes. We
employed feature extraction techniques to identify abnormalities in RPE layer (PED
and CNV) along with the structural and textural properties of the RPE layer (gray
level co-occurrence matrix, GLCM). Specifically, we considered four possible set of
features and for each set we constructed k-NN, naive Bayes, support vector machine
(SVM) and rule-based classifiers. The best classification performance was obtained
for the features capturing the structural and textural properties of the RPE layer
and for naive Bayes classifier (accuracy = 96.1%, sensitivity = 91.3%, specificity =
100.0%) and SVM classifier (accuracy = 94.1%, sensitivity = 100.0%, specificity =
93.7%). Our results confirm the usefulness of the features characterizing the RPE
layer in the diagnosis of wet AMD.
Keywords: wet aged-related macular degeneration, OCT images, gray level co-
occurrence matrix, feature extraction, classification
5.1 Introduction
Age-related macular degeneration (AMD) is one of the major eye-related diseases.

In the United States it affects a huge population every year [1]. AMD is further
specialized into dry AMD and wet AMD. Dry AMD is common in patients aged
Poznan University of Technology, Poznan, Poland

e-mail: anam.haq@put.poznan.pl,szymon.wilk@cs.put.poznan.pl

https://doi.org/10.1007/978-3-319-70063-2_5
44 Anam Haq et al.
above 50 with the possibility that it may lead to the development of wet AMD [2].
Wet AMD is caused by the growth of abnormal blood vessels under the retina. These
blood vessels are fragile and may leak blood or fluid. With the growth up of these
blood vessels, vision becomes blurred and distorted.
Wet AMD in most cases affects the central vision and one can only see using
side vision. Fortunately, it timely detection may save the patient’s normal vision. A
diagnostic process relies on analysis of changes occurring in the retinal pigment ep-
ithelium (RPE) layer. Specifically, to diagnose wet AMD ophthalmologists look for
pigment epithelium detachment (PED) and fluid filled region [2, 3]. The presence
and severity of these pathological changes can be evaluated using optical coherence
tomography (OCT) – a medical imaging technology that is used to diagnose mul-
tiple eye-related diseases [4]. A normal OCT image of the eye macula is shown in
Fig. 41.2.
Fig. 5.1: OCT image: a – a healthy eye, b – an eye with wet AMD
In this paper we present a case study aimed at detecting wet AMD in OCT im-
ages. A considered data set includes 51 images – 21 of wet AMD and 30 of healthy
eyes. From these images we extract structural features related to the abnormalities
(PED, CNV) present in the RPE layer and the features representing quantitative
and textural properties of the RPE layer itself. We group these features into four
sets and for each set we construct several possible classifiers (k-nearest neighbor,
nave Bayes, support vector machine, decision rules). Our goal is to identify a set of
features and a classifier that result in the highest diagnostic accuracy.
5.2 Related Work
OCT is able to visualize the ocular structure at higher resolutions, making it one of
the widely used diagnostic methods [4]. Earlier methods were more time consum-
ing and complex, involving measuring the thickness of receptor layer [3]. Initially,
OCT was only used to obtain the anterior segmentâĂŹs vivo optical cross sections
and retinal diseases, e.g., macular hole, macular detachment, macular edema and
epiretinal membrane [2]. The cross section of OCT images was also used in order to
evaluate the retinal layers like retinal nerve fiber layer and optic disk [5, 6]. There-
5 Detection of Wet Age-related Macular Degeneration in OCT Images: A Case Study 45
fore, the extraction of required segmented information from OCT images became
vital.
The major work regarding OCT is performed in the retinal layers of the eye and
is concerned with the reduction of speckle noise. Speckle noise is a type of gran-
ular noise that inherently exists in the OCT images – it lowers the resolutions and
contrast of the image resulting in its degradation. There are various techniques to
address these issues, e.g., in [7] the authors used wavelet reconstruction and decom-
position to reduce speckle noise. Another important aspect related to the analysis
of retinal layers is the segmentation of PED and CNV in order to detect wet AMD.
Several approaches have been already proposed to to identify these abnormalities.
Fabritius et al. [8] developed a segmentation technique that was based on a vari-
ation of signal intensity. Lee et al. [9] described a method to detect various kinds
of PEDs in OCT images using mean intensity map and standard deviation of the
intensity spread. Haq et al. [10] developed a method combining multiple image pro-
cessing techniques to segment PEDs and CNVs and to extract features capturing
their structural properties. k-means segmentation was used to segment green part
of the OCT as it can be used efficiently in the next step to obtain the structural in-
formation related to PEDs and CNVs. A combination of morphological operations
such as erosion and dilation are performed in a specific sequence to segment the ab-
normalities. In the last step, features consisting of different structural properties are
extracted from the binary image. Prashanth et al. [11] extracted quantitative features
of the RPE layer in OCT images. Finally, Ayaz et al. [12] detection CNVs in the
RPE layer by extracting textural features from that layer.
5.3 Methods
In this section we describe methods employed in our case study. They were orga-
nized into a computational work-flow with the following stages: (1) preprocessing,
(2) segmentation, (3) feature extraction, and (4) classifiers.
5.3.1 Preprocessing
OCT images are basically a combination of red, green and blue color channels as
shown in Fig. 41.2. The blue part of the OCT image does not provide any important
information as it consists of vitreous humour layer of the eye, thus it was treated
as noise and removed using k-means segmentation. Further preprocessing was con-
ducted in two parallel branches. The first branch was concerned with preparing im-
ages for the segmentation of abnormalities (PEDs and CNVs) and it involved the
steps followed presented in [10]. The second branch was aimed at processing the
RPE layer – here red and green segments of the image obtained using k-means seg-
46 Anam Haq et al.
mentation were combined together with averaging, the resulting image was then
converted to grayscale and passed to a median filter.
Fig.33.3a and Fig.33.3b present the results of the second preprocessing branch
for an image of a healthy eye and a sick eye respectively.
5.3.2 Segmentation
Abnormalities like CNVs and PEDs were obtained by applying k-mean segmen-
tation on the images. In the first phase, green segment image was obtained from
segmentation, next a series of morphological operations were performed to extract
the abnormalities (PEDs and CNVs). In the last step features defining the structural
properties of the abnormalities were extracted. This approach is described in [10].
To segment the RPE layer we followed the following process. In the first step the
preprocessed image was passed onto the thresholding block which calculated the
threshold value using the Otsu method [13]. Then, the image was binarized using
the established threshold. In the next step a series of morphological operations (di-
lation, erosion, opening and closing) were performed [14]. After this the connected
component having the largest area (i.e. the RPE layer) was extracted. The extracted
part and the preprocessed grayscale image were multiplied and the resulting image
was used in the next stage for feature extraction.
Results of segmentation of the RPE layer are shown in Fig. 5.2e and Fig. 5.2f.
Fig. 5.2: Results of preprocessing and segmentation for a healthy (upper row) and
sick eye (lower row): a, b – preprocessing, c, d – binarization, e, f – multiplication
5.3.3 Feature Extraction
In this case study we considered 24 features presented in Table 5.1. They provide
structural characteristics of PEDs and CNVs, structural characteristics of the RPE
layer and textural characteristics of the RPE layer.
The structural features of PEDs and CNVs (f1 – f15) were extracted following
the approach proposed in [10]. However, in this work we provided a separate set
of features for PEDs and CNVs (in the initial proposal there were collapsed by
averaging) and we considered more features (previously, only area, orientation and
axes lengths were used). The structural features of the RPE layer (f16 – f20) were
obtained from the segmented RPE layer following the approach from [11]. Finally,
the textural features of the RPE layer (f21 – f24) were obtained by using gray level
co-occurrence matrix (GLCM) [15], these features are extracted from the images
similar to the ones shown in Fig. 33.3e and Fig. 5.2f.
Values of the features were extracted from OCT images, thus each image was
transformed into a numerical vector with 24 elements. These vectors formed an
input to learning algorithms applied in the last stage. In addition to the set of all
features from Table 5.1, we also introduced its three subsets to get better insight into
their diagnostic capabilities: PED+CNV – structural features of PEDs and CNVs,
RPE1 – selected structural and textural features of the RPE layer (f20 – f23) that
were initially suggested in [12], and RPE2 – all structural and textural features of
the RPE layer (a more comprehensive characteristics than RPE1).
5.3.4 Construction of Classifiers
At this stage we used the numerical data constructed in the feature extraction sec-
tion to construct classifiers distinguishing between wet AMD and healthy eyes. We
considered the following classifiers: k-nearest neighbor (k-NN) k = 1 and k = 3
(denoted further as 1NN and 3NN respectively), naive Bayes (NB) with kernel den-
sity estimator, support vector machine (SVM) using linear kernel and decision rules
induced with PART algorithm (PART). We decide to use both non-symbolic and
symbolic classifiers to maximize the classification performance and to get better
insight into the knowledge discovered from data.
The workflow described in Section 5.3 was applied to a set OCT images obtained
from Armed Forces Institute of Ophthalmology (AFIO), Rawalpindi, Pakistan. This
set contains 51 planar (2D) images – 21 show wet AMD, and the remaining 30
represent healthy eyes. Specific planar images were selected from volumetric data
prior to our analysis by clinical experts form AFIO.
48 Anam Haq et al.
Table 5.1: Features considered in our study and their subsets (stdev = standard de-
viation)
Subset
Feature PED+CNV RPE1 RPE2
f1 PED – area
f2 PED – orientation
f3 PED – eccentricity
f4 PED – minor axis length
f5 PED – major axis length
f6 PED – equivalent diameter
f7 PED – mean intensity
f8 CNV – area
f9 CNV – orientation
f10 CNV – eccentricity
f11 CNV – minor axis length
f12 CNV – major axis length
f13 CNV – equivalent diameter
f14 CNV – mean intensity
f15 CNV – maximum intensity
f16 RPE – maximum thickness
f17 RPE – mean thickness
f18 RPE – stdev of thickness
f19 RPE – area
f20 RPE – Euler number
f21 RPE – energy
f22 RPE – homogeneity
f23 RPE – correlation
f24 RPE – contrast
Image processing (stages (1) – (3) in our workflow) were conducted using MAT-
LAB, and the last stage was performed in WEKA. Given a limited number of im-
ages, we employed the leaving-one-out validation schema to evaluate constructed in
the last stage. In this scheme 50 images were used as a learning set and 1 image was
used for testing. This process was repeated 51 times and results were averaged over
repetitions. We consider the following evaluation measures: accuracy, sensitivity
and specificity (assuming the wet AMD class to be the positive one) and geometric
mean (G-mean) of sensitivity and specificity. The latter measure is well suited to
imbalanced data and it allows to avoid the bias associated with over representation
of selected classes [16].
Detailed evaluation results are given in Table 5.2. The most important observa-
tions are the following:
• Using structural and textural properties of the RPE layer resulted in the best
classification performance in terms of accuracy and G-mean, and the structural
features of PEDs and CNVs (PED+CNV) turned out to be less useful. More-
over, with the exception of 1NN, the RPE2 subset worked better than the RPE1
subset proposed in [12],
• For most classifiers using all features deteriorated the classification performance
(with regard to all considered measures) in comparison to the RPE2 subset – this
further emphasizes the usefulness of the latter subset.
• The best performing classifiers were NB and SVM – the former was better
in recognizing healthy eyes (specificity = 100.0%), while the latter correctly
recognized all cases of wet AMD. This finding suggest it may be beneficial to
combine both classifiers to further enhance their diagnostic capabilities.
• Although PART performed worse than the other classifiers, it gave us insight
into the discovered knowledge. For the RPE2 features, the classifier included
only two rules based on the standard deviation of the RPE layer thickness (f18)
and the Euler number (f20), thus confirming the need to consider both structural
and textural features of the RPE layer when diagnosing wet AMD.
Table 5.2: Performance of classifiers using various feature sets (Acc = accuracy,
Sens = sensitivity, Spec = specificity)
Feature set Classifier Acc [%] Sens [%] Spec [%] G-mean
PED+CNV 1NN 88.2 89.5 87.5 88.5
RPE1 1NN 96.1 95.2 96.7 96.0
RPE2 1NN 94.2 95.0 93.5 94.3
All 1NN 92.2 87.0 96.4 91.6
PED+CNV 3NN 84.3 84.2 84.4 84.3
RPE1 3NN 92.2 94.7 90.6 92.6
RPE2 3NN 94.1 95.0 93.5 94.3
All 3NN 92.2 94.4 93.5 94.0
PED+CNV NB 88.2 86.0 90.0 87.8
RPE1 NB 94.1 95.0 93.5 94.3
RPE2 NB 96.1 91.3 100.0 95.5
All NB 92.2 87.0 96.4 91.2
PED+CNV PART 88.2 85.7 90.0 87.8
RPE1 PART 92.2 90.5 93.3 91.9
RPE2 PART 94.2 91.9 96.5 93.7
All PART 88.2 85.7 90.0 87.8
PED+CNV SVM 86.3 88.9 84.8 86.8
RPE1 SVM 78.4 75.0 80.6 77.8
RPE2 SVM 94.1 100.0 93.7 96.8
All SVM 92.2 90.5 93.3 91.9
50 Anam Haq et al.
5.5 Conclusion
In this study we considered different types of features corresponding to the abnor-

malities present in the RPE layer (PEDs and CNVs) along with the characteristic
of the RPE layer. We extracted these features from OCT images and used them to
build several classifiers, including k-NN (with k = 1 and k = 3), naive Bayes, support
vector machine and a rule-based classifier. Classifiers constructed using 8 features
characterizing the RPE layer were most accurate – this finding clearly shows bene-
fits of adding new features to the set of 4 features initially proposed in [12] and also
the usefulness of these features in diagnosing wet AMD. Features related to abnor-
malities turned out to be less useful. Also, using all considered features resulted in
deteriorated performance. The best classification performance was achieved by the
naive Bayes classifier (accuracy = 96.1%, sensitivity = 91.3%, specificity = 100.0%)
and the support vector machine (accuracy = 94.1%, sensitivity = 100.0%, specificity
= 93.7%). Complementary behavior of these two classifiers suggests they may be
combined together for even better diagnostic performance.
Acknowledgments. The authors would like to thank Dr. Muhammad Usman Akram
from National University of Sciences and Technology, Pakistan for the OCT images
and also to the physicians at AFIO, Pakistan for their guidance.
References
1. National Eye Institute: Facts about age-related macular degeneration (2015)

2. Kanski, J.J., Bowling, B.: Clinical Ophthalmology: A Systematic Approach. Elsevier Health
Sciences (2011)
3. Denniston, A.K.O., Murray, P.I.: Oxford Handbook of Ophthalmology. Oxford University
Press (2014)
4. Keane, P.A., Patel, P.J., Liakopoulos, S., Heussen, F.M., Sadda, S.R., Tufail, A.: Evaluation
of age-related macular degeneration with optical coherence tomography. Surv. Ophthalmol.
57(5) (2012) 389–414
5. Walecki, W.J., Szondy, F.: Fiber optics low-coherence IR interferometry for defense sensors
manufacturing. In: Procedings of SPIE. Volume 7322., SPIE (2009)
6. Walecki, W.J., Pravdivtsev, A., Santos, M., Koo, A.: High-speed high-accuracy fiber optic low-
coherence interferometry for in situ grinding and etching process monitoring. In: Procedings
of SPIE. Volume 6293., SPIE (2006)
7. Gupta, V., Chan, C.C., Poh, C.L., Chow, T.H., Meng, T.C., Koon, N.B.: Computerized au-
tomation of wavelet based denoising method to reduce speckle noise in OCT images. In:
International Conference on Information Technology and Applications in Biomedicine, Pro-
cedings of IEEE (2008)
8. Fabritius, T., Makita, S., Miura, M., Myllylä, R., Yasuno, Y.: Automated segmentation of the
macula by optical coherence tomography. Opt. Express 17(18) (2009) 15659–15669
9. Lee, S.Y., Stetson, P.F., Ruiz-Garcia, H., Heussen, F.M., Sadda, S.R.: Automated characteri-
zation of pigment epithelial detachment by optical coherence tomography. Invest Ophthalmol
Vis Sci 53(1) (2012) 164–170
10. Haq, A., Mir, F.J., Yasin, U.U., Khan, S.A.: Classification of Wet Aged Related Macular De-
generation using Optical Coherence Tomographic Images. In: Proceeding of 6th International
Conference on Machine Vision. Volume 9067. (2013)
11. Prashanth, R., Paranjape, S.V., Ghosh, S., Dutta, P.K., Chatterjee, J.: Characterization of
changes in retinal pigment epithelium layer in choroidal neovascularization through analysis
of optical coherence tomographs. In: 2010 IEEE Students Technology Symposium (Tech-
Sym), IEEE (2010) 39–43
12. Ayaz, S., Sahar, S., Zafar, M., Akram, M.U., Nadeem, Y.: Analysis of OCT images for detec-
tion of choroidal neovascularization in retinal pigment epithelial layer. In: Neural Information
Processing: 21st International Conference, ICONIP 2014. Springer (2014) 226–233
13. Otsu, N.: A threshold selection method from gray-level histograms. IEEE Trans. Syst., Man,
Cybern., Syst 9(1) (1979) 62–66
14. Zhao, F., Zhang, J., Ma, Y.: Medical image processing based on mathematical morphology.
In: International Conference on Computer Application and System Modeling, Procedings of
ICCASM (2012)
15. Saroja, G.A.S., Sulochana, C.H.: Texture analysis of non-uniform images using glcm. In:
2013 IEEE Conference on Information Communication Technologies. (2013) 1319–1322
16. Kubat, M., Matwin, S.: Addressing the curse of imbalanced training sets: One-sided selection.
In: Proceedings of the Fourteenth International Conference on Machine Learning, Morgan
Kaufmann (1997) 179–186
Chapter 6
Gender recognition using artificial neural
networks and data coming from force plates
Jakub Krzysztof Grabski, Tomasz Walczak, Martyna Michałowska, and Magdalena

Cieślak
Abstract. The paper deals with a problem of automatic gender recognition based
on parameters obtained from the force plates. The ground reaction force is recorded
and some selected parameters of the curve are calculated. These parameters are used
in this study as inputs to artificial neural network which should recognize if the in-
dividual is male or famale. The results of recognition are satisfactory and presented
in the paper.
Keywords: gender recognition, human gait, ground reaction force, artificial neural
networks
6.1 Introduction
Human gait is one of a basic study in biomechanics. It is defined as a cyclic combi-

nation of movements which results in human locomotion [1]. During the cycle dif-
ferent phases can be distinguished (basic phases are the stance and swing phases).
Human gait is analyzed using specialized equipment, e.g. motion analysis systems,
force plates, electromyography and many others [2]. Using such an equipment dif-
ferent normal and pathological functions of human locomotion can be studied [3].
It is well known that differences between individuals can be observed. Differences
in human gait pattern are observed also between females and males. The differences
result mainly from body building of men and women. The ground reaction force
observed during the human gait is very characteristic. In the paper the data coming
from the force plates are used as the inputs to the artificial neural network in order
to recognize gender.
On the other hand more and more popular is people identification using different
human characteristics. It is called biometrics [4]. One of the oldest and most popu-
Institute of Applied Mechanics, Faculty of Mechanical Engineering and Management, Poznan Uni-
versity of Technology, Piotrowo 3, 60-965 Poznań, Poland
e-mail: jakub.grabski@put.poznan.pl

https://doi.org/10.1007/978-3-319-70063-2_6
54 Jakub Krzysztof Grabski et al.
lar techniques is recognition using fingerprints [5]. However there are many others
techniques of people identification using biometrics, e.g. face recognition [7], hand
geometry [9], palm print [10], DNA [11], palm veins [12] or iris recognition [8].
Gender recognition is a type of so called soft biometrics [6]. In such a case
individuals are grouped into different categories related to their physical traits
(e.g. height, weight, skin or eyes colour, etc.), behavioural traits (such as gait or
keystroke) or adhered human characteristics (e.g. tattoos). Recognition of gender
have many useful potential applications, e.g. human-computer interaction systems,
surveillance systems or counting the number of males and females. It can be used
also in computer-aided physiological or psychological analysis systems. There are
many different techniques of automatic gender recognition in the literature. The
most popular technique is based on visible images of people [14]. However there
are many different problems related to this approach. One of the new approach is
gender recognition using thermal infrared images [15]. Some researchers recognize
gender based on voice [13]. The paper presents an attempt to recognize gender based
on the data from the force plates. Previously similar study with these data has been
conducted to recognize individuals [16, 17].
In the paper parameters obtained from analysis of the ground reaction force dur-
ing single human gait cycle are used as the inputs to the artificial neural network
(ANN). The task of the ANN is to recognize gender based on these parameters.
6.2 Human gait and ground reaction forces
Human gait is a cyclic movement and in general the cycle is divided into the stance
and swing phases. The first one takes about 60% of total time of the cycle and the
second one takes about 40% of the total time.
The stance phase consists of:
• initial contact phase,
• loading response phase,
• mid stance phase,
• terminal stance phase,
• toe off phase.
In a similar way the swing phase can be divided into:
• initial swing phase,
• mid swing phase,
• terminal swing phase.
From the other hand the cycle of the human gait can be also devided into the
single support and double support phases. During the double support phase both
legs are in contant with the ground. The ground reaction force (GRF) for both legs
takes non-zero values. In the single support phase only one leg remain in contant
with the ground.
6 Gender recognition 55
Human gait can be analyzed by observing the GRF. Fig. 1 shows vertical compo-
nent of the GRF for the left and right legs. For this graph the following parameters
are defined in this study:
• instant of the maximum of the loading response phase (parameter 1),

• maximum of the loading response phase (parameter 2),
• instant of the minimum of the midstance phase (parameter 3),
• maximum of the terminal stance phase (parameter 4),
• instant of the maximum of the terminal stance phase (parameter 5),
• minimum of the midstance phase (parameter 6),
• maximum of the loading response phase divided by minimum of the midstance
phase (parameter 7),
• maximum of the terminal stance phase divided by minimum of the midstance
phase (parameter 8),
• maximum of the loading response phase divided by maximum of the terminal
stance phase (parameter 9),
• period between the minimum of the midstance phase and the maximum of the
loading response phase (parameter 10),
• period between the maximum of the terminal stance phase and the maximum of
the loading response phase (parameter 11),
• area under curve from beginning to the maximum of the loading response phase
(parameter 12),
• area under curve between maxima (parameter 13),
• area under curve from the maximum of the terminal phase to the end of the
stand phase (parameter 14),
• area under curve from beginning to the maximum of the loading response phase
divided by area under curve between maxima (parameter 15),
• area under curve between maxima divided by area under curve from the maxi-
mum of the terminal phase to the end of the stand phase (parameter 16),
• stepness of the curve between the maxima (parameter 17),
• number of the artifacts occuring on the curve in different phases (parameters
18-20),
• width of the artifact on the rising slope (parameter 21),
• height of the artifact on the rising slope (parameter 22),
• instant of the maximal artifact occuring on the rising slope (parameter 23),
• width of the artifact on the falling slope (parameter 24),
• coordinates of the geometrical center of the curve (parameters 25-26),
• time of contant with the ground (parameter 27),
• time of the double support phase (parameter 28).
Fig. 2 presents anterior-posterior component of the GRF for the left and right legs
during single gait cycle. Some characteristics parameters of this graph are defined
in this study:
• instant of the maximum (parameter 29),
• maximum of the GRF anterior-posterior component (parameter 30),
Fig. 6.1: Vertical component of the ground reaction force for the left and right legs
during single gait cycle
• instant of the minimum (parameter 31),

• minimum of the GRF anterior-posterior component (parameter 32),
All of the aboved mentioned parameters were calculated separately for each leg.
Detailed description of these parameters can be found in [16, 17]. The parameters
were calculated using self-implemented methods in MATLAB software.
6.3 Data from the force plates
The data used in the paper were recorded for 15 volunteers, seven females and eight
males. They walked through a walkway with two build-in force plates. During the
walk the subjects put the first leg on the first force plate and the second one on the
second force plate. In such a way a three-dimensional GRF during single gait cycle
was recorded. Below one can find some information about the subjects (average ±
standard deviation and ranges - min ÷ max).
Women:
• age: 24 ± 3.9 (21 ÷ 33) years,
• height: 168.6 ± 3.8 (165 ÷ 176) cm,
• body mass: 60.1 ± 6.9 (54.6 ÷ 76) kg.
Men:
• age: 25 ± 9.9 (18 ÷ 51) years,
• height: 181.6 ± 3.9 (176 ÷ 190) cm,
Fig. 6.2: Anterior-posterior component of the ground reaction force for the left and
right legs during single gait cycle
• body mass: 83.3 ± 14 (69 ÷ 113) kg.

The force plates used in this study were AMTI BP400600. The sampling fre-
quency was equal to 400 Hz.
6.4 Artificial neural network
Fig. 3 presents schematically structure of the artifial neural network (ANN) applied
in the paper. It consists of three layers: two hidden layers and output layer. There are
15 neurons in each hidden layer and one neuron in the output layer. The activation
functions are in the hidden layers and linear function in the output layer. In order to
decide if the individual is recognized as male or female in the end of this structure
the step function is used. The parameters presented in previous section are used as
the inputs to the ANN. The Levenberg-Marquardt method is used in the learning
process. All calculations were conducted in Matlab NN Toolbox.
6.5 Results
The parameters presented in Section 2 were divided into smaller groups:

• group 1: parameters 1-6,
Fig. 6.3: Architecture of the artifial neural network applied in the paper
• group 3: parameters 10-11 and 27-28,

• group 4: parameters 12-16 and 25-26,
• group 6: parameters 29-32.
One sample of data are parameters for a single gait cycle. The data obtained for
15 individuals were divided into the learning and test sets in such a way that the
training set included all data obtained for 11 individuals and the test set included the
remaining data for 4 individuals (2 females and 2 males).
After the learning process accuracy of gender recognition using the ANN is ex-
aminated for the test set using the recognition rate defined as follows
NPR
RR = · 100% (6.1)
TN
where NPR denotes number of properly recognized samples and TN is the total num-
ber of samples in the test set. Tab. 6.1 presents the best results of gender recognition
using the ANN described in the previous section for different combinations of inputs
data.
One can observe that quite high rate of recognition was obtained in this study.
In case of taking all presented here parameters as inputs to the ANN the obtained
recognition rate was 89.02%. Comparing results obtained for single groups of pa-
rameters the best results was obtained for groups of parameters 3-5. One can observe
that also for different combinations of inputs parameters the best results were ob-
tained for inputs including parameters from groups 3-5. The recognition rate was
also examined for combination of these three groups of parameters. The obtained
recognition rate in such a case was 94.03%.
Table 6.1: Results of gender recognition
results for whole only only

inputs from test set female male
all groups 89.02% 79.54% 99.01%
group 1 78.52% 72.09% 85.22%
group 2 58.71% 51.16% 66.50%
group 3 89.26% 90.23% 88.18%
group 4 87.11% 82.33% 92.12%
group 5 84.25% 74.88% 94.09%
group 6 53.94% 43.72% 64.53%
groups 1+2 79.48% 73.49% 85.72%
groups 1+3 78.52% 73.49% 83.74%
groups 1+4 80.90% 71.16% 91.13%
groups 1+5 90.21% 86.98% 93.60%
groups 1+6 74.22% 65.58% 83.25%
groups 2+3 84.49% 77.68% 91.63%
groups 2+4 84.25% 71.63% 97.54%
groups 2+5 89.74% 82.79% 97.04%
groups 2+6 63.72% 52.10% 75.86%
groups 3+4 88.07% 81.86% 94.58%
groups 3+5 88.07% 83.26% 93.10%
groups 3+6 80.67% 77.68% 83.74%
groups 4+5 92.84% 94.88% 90.64%
groups 4+6 85.92% 86.51% 85.22%
groups 5+6 78.76% 70.23% 87.69%
groups 3+4+5 94.03% 95.35% 92.61%
6.6 Summary
In the paper gender recognition has been conducted using the data coming from the
force plates and the ANN. The presented results show quite good accuracy of the
gender recognition rate, in particular for parameters 10-28 taken as inputs to the
artificial neural network.
Acknowledgements
The presented research were funded with grant for science allocated by the Ministry
of Science and Higher Education in Poland (02/21/DSPB/3493).
References
1. Inman, V.T., Ralston, H.J., Todd, F.: Human Walking, Williams & Wilkins (1981).
2. Payton, C.J., Barlett, R.M. (eds.): Biomechanical Evaluation of Movement in Sport and Exer-
cise, Routledge. Taylor & Francis Group, London and New York (2008).
3. Perry, J., Burnfield, J.M.: Gait Analysis. Normal and Pathological Function, SLACK Incorpo-
rated (2010).
4. Anil, J., Ross, A.A., Nandakumar, K.: Introduction to Biometrics, Springer (2011).
5. Maltoni, D., Maio, D., Jain, A., Prabhakar, S.: Handbook of Fingerprint Recognition, Springer
(2009).
6. Dantcheva, A., Velardo, C., D’Angelo, A., Dugelay, J.-L.: Bag of Soft Biometrics for Person
Identification: new trends and challenges, Multimed Tools Appl 51, 739ÂŰ777 (2010).
7. Karczmarek, P., Kiersztyn, A., Pedrycz, W., Dolecki, M.: An application of chain code-based
local descriptor and its extension to face recognition, Pattern Recogn. 65, 26-34 (2017).
8. Thalji, Z., Alsmadi, M.: Iris recognition using robust algorithm for eyelid, eyelash and shadow
avoiding, World Appl Sci 25, 858-865 (2013).
9. Mathivanan, B., Palanisamy, V., Selvarajan, S.: A hybrid model for human recognition system
using hand dorsum geometry and finger-knuckle-print, Journal of Computer Science 8, 1814-
1821 (2012).
10. Steffi Vanthana, P., Muthukumar, A.: Multimodal biometrics authentication using iris and
palmprint with SVM classifier, International Journal of Applied Engineering Research 10,
16271-16277 (2015).
11. Zhou, P., Tian, F., Ren, Y., Shang, Z.: Systematic classification and analysis of themes in
protein-DNA recognition, J Chem Inf Model 50, 1476-1488 (2010).
12. Yan, X., Kang, W., Deng, F., Wu, Q.: Palm vein recognition based on multi-sampling and
feature-level fusion, Neurocomputing 151, 798-807 (2015).
13. ÂŇwiebocka-WiÃłk, J.: Gender recognition based on speakerÂŠs voice analysis, Adv Intel
Syst Comput 539, 80-85 (2017).
14. Damayanti, F., Rachmad, A.: Recognizing gender through facial image using Support Vector
Machine, J Theor Appl Inf Technol 88, 607-612 (2016).
15. Wang, S., Gao, Z., He, S., He, M., Ji, Q.: Gender recognition from visible and thermal infrared
facial images, Multimed Tools Appl 75, 8419-8442 (2016).
16. Walczak, T., Grabski, J.K., Grajewska, M., Michałowska, M.: Application of artificial neural
networks in man’s gait recognition, In: Advances in Mechanics: Theoretical, Computational
and Interdisciplinary Issues. Proceedings of the 3rd Polish Congress of Mechanics (PCM) and
21st International Conference on Computer Methods in Mechanics (CMM), Kleiber, M., Bur-
czyński, T., Wilde, K., Górski, J., Winkelmann, K., Smakosz, Ł. (eds.), CRC Press, Taylor &
Francis Group, London (2016), 591-594.
17. Walczak, T., Grabski J.K., Cieślak M., Michałowska M.: The recognition of human by the
dynamic determinants of the gait with use of ANN. In: Springer Proceedings in Mathematics
and Statistics 181, Dynamical Systems: Modelling, Awrejcewicz, J. (ed.), Springer (2016), 375-
385.
Chapter 7
Human Sperm Morphology Analysis using
a Digital Holographic Microscope
Emil Fabian1 , Marzena Kamieniczna2 , Maciej Kurpisz2 , and Ewa Stachowska1
Abstract. We used a digital holographic microscope (DHM) to analyse human

sperm morphology. Length, width and thickness of the sperm cell head were deter-
mined by an operator, using the images produced by the software provided with the
DHM. We developed an algorithm to replace part of the activities of the operator.
This algorithm was written in the programming language R and the package EBIm-
age. The algorithm produced similar results, while reducing the time needed from
six hours to two when using a standard procedure.
Keywords: digital holography, human sperm, morphology, image analysis
7.1 Introduction
Human sperm parameters such as morphology, number of spermatozoa and pro-

gressive motility are of a great importance to assess male reproductive health. Mor-
phology data are highly predictive for male fertility and are important in selecting
spermatozoa for assisted reproductive techniques (ART). Because decreasing fertil-
ity due to poor semen morphology has been continously observed in men [1], there
is growing interest to understand morphological alterations of spermatozoa. Abnor-
mal sperm morphology — teratozoospermia — means that the percentage ofnormal
spermatozoa in the sample is below 4%, according to Kruger’s strict criteria [2].
To identify morphology by this method 200 spermatozoa are needed. Morphology
is very complex and difficult to investigate. Classical evaluation of sperm morphol-
Department of Metrology and Measurement Systems, Institute of Mechanical Technology, Faculty

of Mechanical Engineering and Management, Poznan University of Technology, ul. Piotrowo 3,
e-mail: ewa.stachowska@put.poznan.pl · Institute of Human Genetics Polish Academy
of Science, ul. Strzeszyńska 32,

https://doi.org/10.1007/978-3-319-70063-2_7
62 Emil Fabian et al.
ogy requires the preparation of smears, fixation and staining of smears. Negative
aspects of this method are: over-staining, artifacts from smearing, staining and/or
fixation [3]. To avoid any alteration of the vitality of the sperm analysed and to
exclude adverse effects that may be caused by labels a non-destructive method is
required, using pure sperm. Such methods use the fact, that spermatozoa are essen-
tially transparent and have a refractive index different from the surrounding medium
[4]. Therefore the phase of the light transmitted by the sample is modulated. A qual-
itative visualization of this phase contrast can be obtained by contrast interference
microscopy. But to produce a 3D-image of a biological sample, this has to be moved
along the z-axis in very small steps to collect images in different focal planes. This
is a difficult and time-consuming method for obtaining quantitative morphological
imaging [5].
Contrary to other methods digital holographic microscopy (DHM) offers label-free,
non contact, non-invasive 3D-measurements with nanometre axial resolution in real
time. This method records and reconstructs the phase and amplitude of the wave-
front transmitted by the sample. An important advantage is that 3D quantitative
sample imaging can be done by digitally reconstructing 2D images at different ob-
ject planes (this is in fact refocusing) without mechanically refocusing [6, 7, 8]. 3D
information is obtained with a single exposure. Adding the third dimension pro-
vides a better understanding of the sperm behaviour and its relation with male in-
fertility [9, 10]. Recently several groups started therefore to use digital holographic
microscopy to analyse human sperm [4, 11, 12, 13, 14, 15, 16, 17]. No full quan-
titative analysis was however made of the morphology and motility of sperm cells
[10]. Our final goal is to determine the morphology and motility of a sperm cell
simultaneously. We developed and verified a simple automatic method to obtain 3D
morphological sperm cell parameters of untainted sperm samples using a DHM, not
dependent on the experience of an operator. For this we use and develop open source
programs, written in the programming language R [18]. In this work we present our
first results.
7.2 Materials and apparatus
7.2.0.1 Materials
We investigated untainted human sperm of fertile and infertile patients, which were
invited to take part in our study. Single semen specimens from two fertile and
one infertile men were studied. Measurements of 200 spermatozoa of each pa-
tient were performed in accordance with the guidelines of the Word Health Orga-
nization (WHO)[2]. To asses the morphometric characteristics of spermatozoa wet
mounted slides from fresh ejaculates were analysed for head length, width and high.
A 10 μ l fresh semen sample was loaded onto a clear slide glass and covered with
a 22x22 mm2 cover glass. The final sperm concentration should be no more than
7 Human Sperm Morphology Analysis using a Digital Holographic Microscope 63
40x106 /ml. If the sperm concentration was higher we decided to dilute the semen
samples by seminal plasma.
7.2.0.2 Apparatus
We used a Lyncée Tec DHM T1000 digital holographic transmission microscope

in an off-axis Mach-Zehnder configuration [19]. The light source is a laser diode
at 666 nm with optical fiber coupling for enhanced stability, illuminating the sam-
ple with very low power (down to 1 μ W/cm2 ). The sperm samples were observed
with a magnification of 50x, using a dry objective/condensor assembly (NA=0.75,
WD=0.37 mm, FOV= 104 μ m). The hologram is registered with a CCD camera
with 1024x1024 pixels at 30 fps. We can obtain both phase and intensity images,
contrary to traditional microscopy where only an intensity image is recorded (see
figure 7.1). From a single recorded hologram, the phase of the image of the sam-
Fig. 7.1: The intensity image (left) and phase image (right) of the sperm samples;
the area of each image is 104 μ m x 104 μ m
ple can be reconstructed numerically [7]. The Lyncée Tec DHM is delivered with
the Koala software, which allows measurement and also interpretation of the data
obtained. Scientific publications can be found at the Lyncée Tec webside [19]. Be-
cause 2D images are numerically reconstructed at different focal planes without
mechanically adjusting the optical system we can measure sperm cells swimming
at different depths in the sample at the same moment (see figure 7.2). On the left
part of figure 7.2 one can see a sharp image of two sperm cells (marked in a full line
oval), which became blurred after numerical refocusing at a different depth (right
part of this figure). The opposite situation can be observed on the same pictures
for the sperm cell out of focus in the left picture and focused in the right (marked
in a dashed oval). It is also possible to extract a profile along a chosen line from
Fig. 7.2: The phase images reconstructed at different focal planes; the area of each
image is 104 μ m x 104 μ m
the phase image (1D-profile, see figure 7.3). This profile we used to measure the
morphological parameters of the sperm head : length, width and thickness.
Fig. 7.3: Screen dump from a window showing a 1D–profile
7.3 Method
We used the data (images) obtained by the KOALA-software to determine the size
of a sperm cell. Two different procedures were used. In the first we choose a line
along which a profile in the picture is defined (see figure 7.3). If this profile is chosen
lengthwise we can determine the length. When chosen crosswise along a trajectory
where the operator estimates the maximum width, we obtain the width of the sperm
cell. Both profiles provide also the height. From these both profiles together we can
estimate the volume as well.
A second approach was to cut from a phase image a region of interest (ROI),
showing a clear sperm cell, not disturbed by another one. Then the background was
removed and thresholding applied. Thereafter routines from the R–package EBIm-
age [18, 20, 21] were used to determine length and width of the selected sperm
cell. The thresholded image (see figure 7.4) is used as a contour to cut the image
of the sperm cell out of the region of interest of the phase–image. The intensity in
the phase–image in the Koala software is a measure of the thickness of an object.
The volume can now easily be obtained by summing all the pixels in the ROI; the
background after thresholding is by definition 0. The program we wrote calculated
from an image of a ROI the three parameters: length, width and thickness.
Fig. 7.4: A ROI with one sperm cell; left: the original phase–image (y); right:
the thresholded image (x); centre: the product of the two (y∗x)
7.4 Measurements and results
The first procedure described in 7.3 was used by two operators, using each the same
50 sperm cells. The results are given in table 7.1. It is clear that the influence of the
operator can be neglected.
To compare both procedures the sperm of two healthy persons and one infertile
was measured by an experienced operator and also determined using our analysis
software. The results for 200 sperm cells of each patient are presented in table 7.2.
In this case the selections of sperm cells used by the operator and the algorithm
were not identical. The results are in good agreement, except for the thickness. The
software finds the maximum thickness considering the whole sperm cell as indi-
cated in figure 7.4 (centre). The operator has to choose a cross-section in figure 7.1
(right) and determine the thickness from this profile. It is very difficult to select the
Table 7.1: Comparison of measurements by two operators (Marzena Kamieniczna–

MK resp. Emil Fabian–EF) on 50 sperm cells
operator MK EF
Length, Width L W T L W T
Thickness [μ m] [μ m] [μ m] [μ m] [μ m] [μ m]
average value 5.0 2.84 1.89 5.1 2.90 1.93
2σ 1.2 0.52 0.47 1.0 0.60 0.48
min value 3.81 2.41 1.41 4.12 2.24 1.42
max value 6.40 3.90 2.83 6.15 3.81 2.84
range 2.59 1.49 1.42 2.03 1.57 1.42
Table 7.2: Comparison of measurements by an operator and our algorithm on sam-

ples of 200 sperm cells
operator MK algorithm
Length, Width L W T L W T
Thickness [μ m] [μ m] [μ m] [μ m] [μ m] [μ m]
fertile patient 1
average value 5.2 3.1 1.95 5.5 3.0 2.12
2σ 1.2 0.8 0.65 1.5 0.6 0.35
fertile patient 2
average value 5.3 3.1 1.95 5.2 2.59 1.95
2σ 1.2 0.8 0.65 1.2 0.51 0.30
infertile patient 3
average value 4.9 3.39 1.89 5.7 2.9 2.10
2σ 1.1 0.73 0.72 1.9 0.97 0.35
profile with the largest thickness. Therefore the values for the thickness found by an
operator are systematically lower. But it is clear that our algorithm produces simi-
lar values and gives even better results for the thickness. It saves furthermore much
time, using only two hours in stead of six per patient.
We found different shapes of the distributions of the head lengths of sperm cells
for a fertile and a non-fertile patient (see figure 7.5). A shape similar to the distri-
bution in this figure for a fertile patient was also found for the other fertile one. We
obtained similar results for the widths and the thicknesses. This indicates that the
shape difference could be used as a diagnostic tool. We will investigate this for more
patients.
Comparison of sperm cell morphology with results obtained with digital holo-
graphic microscopy found in literature [4, 15, 22, 23] is not well possible, even when
Fig. 7.5: Histograms of the head lengths of sperm cells of a fertile patient (left) and
a non-fertile one (right)
the results are largely similar, because we used live, non-labelled, non-prepared
sperm cells swimming in the semen of the patient.
7.5 Conclusions
Our first results show that our algorithm gives data comparable with those obtained
by an operator, while saving a large amount of time. We are also optimising our
program as well as decreasing the need for operator activity. The method can be
extended to determine other morphological parameters and in particular to study the
existence of vacuoles in the sperm head.
Acknowledgements
EF and ES gratefully acknowledge support by the Poznan University of Technology

under project nr 02/22/DSPB/1387.
We thank dr F.G. Meijer for discussing the manuscript.
References
1. Kruger, T.F., DuToit, T.C., Franken, D.R., Acosta, A.A., Oehninger, S.C., Menkveld, R., Lom-
bard, C.J. : A new computerized method of reading sperm morphology (strict criteria) is as
efficient as technician reading. Fertil. Steril. 59(1), 202–209 (1993)
2. Word Health Organization : Reference values and semen nomenclature. In WHO Laboratory
Manual for the Examination and Processing of Human Semen. 5th ed., 223–225 Geneva: WHO
Press (2010)
3. Sikka, S.C., Helstorm, W.J.G. : Current updates on labolatory techniques for the diagnosis of
male reproductive failure. Asian Journal of Andrology, 18,392–401 (2016)
4. Coppola, G., Di Caprio, G., Wilding, M., Ferraro, P., Esposito, G., Di Matteo, L., Dale, R.,
Coppola, G., Dale, B. : Digital holographic microscopy for the evaluation of human sperm
structure. Zygote 22(4), 446–454 (2014)
5. Pieper, R.J., Korpel, A. : Image processing for extended depth of field. Appl. Opt. 22, 1449–
1453 (1983)
6. Schnars, U., Juptner, W. : Direct recording of holograms by a CCD target and numerical recon-
struction. Appl. Opt. 33, 179–181 (1994)
7. Cuche, E., Marquet, P., Depeursinge, C. : Simultaneous amplitude-contrast and quantitative
phase-contrast microscopy by numerical reconstruction of Fresnel off-axis holograms. Appl.
Opt. 38, 6994–7001 (1999)
8. Schnars, U., Juptner, W. : Digital recording and numerical reconstruction of holograms. Meas.
Sci. Technol. 13, R85–R101,(2002)
9. Guerrero, A., Carneiro, J., Pimentel, A., Wood, C.D., Corkidi, G., Darszon, A. : Strategies for
locating the female gamete: the importance of measuring sperm trajectories in three spatial
dimensions. Mol. Hum. Reprod. 17, 511–523 (2011)
10. Di Caprio, G., Ferrara, M.A., Micco, L., Merola, F., Memmolo, P., Ferraro, P., Coppola, G. :
Holographic imaging of unlabelled sperm cells for semen analysis: a review. J. Biophotonics
8(10), 779–789 (2015)
11. Mico, V., Zalevsky, Z., Ferreira, C., Garcia, J. : Superresolution digital holographic mi-
croscopy for three dimensional samples. Opt. Exp. 16, 19260–19270 (2008)
12. Coppola, G., Di Caprio, G.,Gioffré, M., Puglisi, R., Balduzzi, D., Galli, A., Micco, L., Paturzo,
M., Grilli, S., Finizio, A., Ferraro, P. : Digital self-referencing quantitative phase microscopy
by wavefront folding in holographic image reconstruction. Opt. Lett. 35, 3390–3392 (2010)
13. Micco, L., Finizio, A., Puglisi, R., Balduzzi, D., Galli, A., Ferraro, P. : Dynamic DIC by digital
holography microscopy for enhancing phase-contrast visualization. Biomed. Opt. Express 2(2),
331–344 (2011)
14. Memmolo, P., Di Caprio, G., Distante, C., Paturzo, M., Puglisi, R., Balduzzi, D., Galli, A.,
Ferraro, P. : Morphological analysis framework of living cells by digital holography. Opt. Exp.
19, 23215–23226 (2011)
15. Crha, I., Zakova, J., Huser, M., Ventruba, P., Pohanka, M. : Digital holographic microscopy in
human sperm imaging. J. Assist. Reprod. Genet. 28, 725–729 (2011)
16. Iglesias, I., Vargas-Martin, F. : Quantitative phase microscopy of transparent samples using a
liquid crystal display. J. Biomedical. Optics 18(2), 026015 (2013)
17. Girshovitz, P., Shaked, N.T. : Doubling the field of view in off-axis low-coherence interfero-
metric imaging. Light Sci. Appl. 3, e151:DOI:10.1038/lsa.2014.32 (2014)
18. https://www.R-project.org/ (accessed 22 May 2017)
19. https://www.lynceetec.com/ (accessed 20 May 2017)
20. https://bioconductor.org/packages/release/bioc/html/EBImage.html/ (accessed 22 May 2017)
21. Pau, G., Fuchs, F., Sklyar, O., Boutros, M., Huber, W. : EBImage–an R package for image
processing with applications to cellular phenotypes. Bioinformatics 26, 978–981 (2010)
22. Haifler, M., Girshovitz, P., Band, G., Dardikman, G., Madjar, I., Shaked, N. T. : Interferometric
phase microscopy for label–free morphological evaluation of sperm cells. Andrology 104, No.
1 43–47 (2015)
23. Coppola, G., Di Caprio, G., Wilding, M., Ferraro, P.,Esposito, G.,DiMatteo, L., et al. : Digital
holographic microscopy for the evaluation of human sperm structure. Zygote 22, 446–454,
(2014)
Chapter 8
Idea and measurement methods used in
bioimpedance spectroscopy.
Barbara Szuster, Zbigniew Szczurek Dawid Roj, Pawel Kowalski, Aleksander

Sobotnicki, and Jakub Woloszyn
Abstract. Increasing popularity of devices used to measure body composition and

the competition on that specific sector of the market resulted in increased interest
in this subject matter. There are more and more articles appearing that treat of that
subject and construction design descriptions. The aim of this article is to present
the nature of the impedance measurement of the human body and to present various
measurement methods, focusing especially on the electrical bioimpedance measure-
ment. The article describes the device used for body composition measurement, that
was created in the ITAM in Zabrze. Also, a short review of bioimpedance measure-
ment devices has been made, describing their advantages and disadvantages, along
the description of other measurement methods.
Keywords: bioimpedance, BIS, BIA, DXA, body composition, impedance of body,

spectroscopy
8.1 Introduction
The electrical properties of tissues have been known for over 50 years. The two
electrode body water measurement was described in the first half of 20th century
by Barnett and Thomasset. Later, four electrode method was described by Hoffer
and his associates. In the 70s Nyober along with his associates have found a con-
nection between the measured body impedance variances and pulsatory blood flow
in the intestines, pulse itself and breathing. The discovery of electrical properties
of the human body was a breakthrough in creation of non-invasive techniques for
morphological body composition assessment. The production of devices used for
the measurement of body impedance (BIA - bioelectrical impedance analysis) is
based on the measurement techniques using the low current intensity at very high
ITAM, Roosevelta 118, Zabrze, Poland,

e-mail: basiasz@itam.zabrze.pl,
WWW home page: www.itam.pl

https://doi.org/10.1007/978-3-319-70063-2_8
70 Barbara Szuster et al.
frequency. Analyzers allow you to estimate parameters such as body fat content, as
well as body water content, distinguishing between intra- and extracellular water
[3]. Those devices allow for accurate, reliable, safe and noninvasive measurement,
for both healthy people and those that have illnesses like diabetes, hypertension,
obesity and other. The aim of this article is to show basics of the biophysical analy-
sis of the body composition, used measurement methods, devices used for the body
composition measurement and the course of action during the examination.
8.2 Impedance measurements - physical basis
Bioimpedance measurements are based on measurement of the average resistance

of the whole body, being the fusion of resistance (active resistivity) and reactance
(passive resistivity). Impedance value is described for serial connection model by
following equation.

Z = R2 + X 2 (8.1)
Reactance Xc (passive resistivity) is reciprocal of pulsation (circular per angular
frequency) omega of alternating current flowing through the the capacitive element
C of electrical circuit.
1
Xc = [Ω ] (8.2)
ωC
To estimate the human body impedance, one should treat it as serial connection
of resistance and capacitance. Therefore for the description of the parallel circuit
of RC components physical value called the admittance (reciprocal of impedance)
should be used.

1 1
Y= + (8.3)
R2 Xc2
Xc R
Z= (8.4)
R2 + Xc2
Relationship between resistance R and reactance Xc after it’s plit on complex
plane is defined as phase angle.
Xc
Z = − arctan (8.5)
R
For healthy people and the current frequency at 50kHz the angle is between 8 and
15 degrees. The value of the angle gets bigger along with the increasing amounts of
cell membranes, and decreases when there’s increasing amount of fat, proteins and
water with electrolytes. Fat tissue and extracellular water do not exhibit reactance
characteristics (resistivity capacitance), because they not behave as capacitors, al-
8 Bioimpedance spectroskopy 71
though they possess active electrical resistivity (resistance). Whereas reactance cre-
ated on the cell membrane tissue with high amount of water, works like capacitor,
made from two layers (phospholipid hydrophilic fragments that conduct the cur-
rent on the outside and inside the cell membrane). The 50kHz current frequency
is also called "medical frequency". It is the current with intensity lower than 1A
and 50kHz frequency that bypasses the cell internals (only a fraction passes through
them) while the current with frequency greater than 100kHz penetrates cells and in-
tracellular spaces. As the frequency goes above 100kHz, cells with the surrounding
membrane lose their dielectric properties and acquire properties of conductors. It
should also be noted that the impedance decreases with the increasing frequency.
The impedance also depends on the sex and state of health. Women have higher
body impedance than men, which is caused by the presence of bigger amount of
subcutaneous fat and lower density of sweat glands. People with vitamin A defi-
ciency or hypothyroidism have a higher body impedance value, caused by dryness
of the skin and callous layer of the skin. On the other hand, hyperthyroidism causes
a reduction of skin impedance [4,5,6].
8.3 Electrical bioimpedance
The division of measurement methods can be done in two ways. The first would
be in terms of the number of electrodes used in the measurement and the second
with regard to the number of frequencies used in the measurement. According to
the second criterion, the most simple method that uses only one frequency is the
SF-BIA method. It’s based on the measurement with one, 50kHz frequency and it is
widely used in different types of bathroom scales. The next method is based on the
impedance measurement with few frequencies, hence the name multiple frequency
bioimpedance analysis (MF-BIA). Another method allows for the measurement in a
very wide spectrum range, so the measurement would allow to reconstruct the whole
characteristic of the Cole Cole model. The methods can also be divided in regards
of how the patient is situated during the measurement and the type of the devices
used for the measurement - horizontal or vertical. / and the way the measurement
is done - vertically or horizontally. Vertical devices have 8 electrodes, two pairs
are placed on the hands and remaining two pairs are placed on the feet. During the
examination it is very important, that the limbs can’t touch each other. Some of the
devices available on the market allow for four electrode measurement of a certain
part of the body. The method is based on two current electrodes and two voltage
electrodes. They are placed on the segment of the body that has to be measured, eg.
on the chest and abdomen[5].
The correctness of the measurement is influenced by the conditions in which the
study was carried out. It is important to:
• not drink the coffee, alcohol, energy drinks, work out extensively, take hot baths
and take any massages 12 hours before the examination,
• be examined on empty stomach or at least 3-4 hours after the last meal, because
the meal increases the measured impedance value,
• 30 minutes before the exam one should urinate,
• 5 - 10 minutes before examination one must acquire a body posture required
during examination,
• providing accurate weight and height values is critical for getting proper results,
• skin that the electrodes will be put on, has to be cleaned and degreased with
alcohol,
• upper limbs cannot come into contact with hips and lower parts of the body,
• upper and lower limbs should be kept up straight.
The most accurate measurements are obtained by repeating the examination at

the same time of day. It should also be mentioned that there is a method that’s not
based on the bioimpedance measurement but on the bioimpedance X-ray imaging.
It is a method that originally used for assessment the corpse composition, but now
is eagerly used by doctors to measure the living body composition. Also it is known
as the "gold measurement standard" due to the high repeatability and accuracy of
the measured values.
8.4 Body impedance device developed at ITAM
In the Institute of medical technology and equipment ITAM in Zabrze a device used
for the measurement bioimpedance spectroscopy has been created. It has the fol-
lowing dimensions: 120x75x25. It has the ability to communicate with the devices
via Bluetooth 2.1 for wireless data transmission and control. The device has been
portrayed on the image below:
The device is composed from few modules: Bluetooth, charging/discharging sys-
tem for the battery, bioimpedance measurement system, main control unit and op-
erator interface. There’s a button in the middle of the casing used for turning the
device on and off. There are also 2 diodes that are part of the interface. The first,
yellow LED, indicates the battery level, whether it is charged or not and if the de-
vice is turned on. The second, blue LED, is used for indicating whether the device
has established a connection and if it communicates with the monitoring station.
This device allows for the bioimpedance measurement of any segment of the body
with the tetra polar method in the 0 to 1000 om range. Measurements carried out
for the frequencies from 5kHz to 100kHz with a 5kHz step. This approach allows to
extrapolate the bioimpedance spectroscopy measurement to the whole range, that is
from 0 to infinity. This device can cooperate with any PC-class machine that has a
specialized application installed. The application can determine things such as:
• fat free mass FFM,
• fat mass FM,
• total body water TBW,
• intracellular water ICW,
Fig. 8.1: Device for monitoring bioimpedance - ITAM.
• extracellular water ECW.

An additional objective of this program is to send and receive orders between the
device and the supervisor. The supervising device may be a personal computer, but
it also could be a smartphone or a tablet. The system allows for data archiving - it
can store, among other things, patient’s data and test results.
8.5 Comparison of different commercially available devices for

bioimpedance measurements
On the market there are a lot of devices that allow for the measurement of body
composition. They are briefly described below. The comparison of this devices can
be done in various ways, for example looking from the functionality point of view
and technical parameters. Below you can find a comparison based on those two
criteria. The following list is based on the technical parameters criteria.
The technical criteria division can also be made in another way. For example, it
can be divided into stationary and mobile device. Stationary devices usually offer
automatic weight measurement, while the mobile one demands that you enter it after
you measure it by yourself. Stationary devices are powered from a wall socket, are
bigger and heavier than the mobile ones, but their advantage is the lack of need for
charging or replacing the batteries. Division can also be made for those that offer the
whole body measurement or the ones that can measure its segments. Stationary units
usually allow the measurement of the whole body and the electrodes are arranged
in a special rack. Below you can find two examples of stationary devices.
Table 8.1: Comparison of devices for bioimpedance measurement.
Bodystat 1500 Akren BIA101 IMPEDIMED Seca mBCA 515 Tanita MC 780
Anniversary SFB7
sport edition
Measurement BIA BIA BIS BIS BIA
technology
Measurement 20-1300 0-999 50-1100 10-1000 75-1,5
range
Measurement 50kHz 50kHz 256 measure- 1; 1,5; 2; 3; 5; 5kHz; 50kHz;
frequencies ment 4kHz - 7,5; 10; 15; 20; 250kHz
1000kHz 30; 50; 75; 100;
150; 200; 300;
500; 750; 1.000
kHz
PC Connectivity Bluetooth USB cable Ethernet, in- USB, bluetooth USB
tranet
Power source 6×battery AA 2×accumulator Li-Ion acumula- Power supply Power supply
1,5 V tor
Dimensions 185 × 115 × 38 200 × 150 × 65 190 × 130 × 110 976 × 1251 × 450 × 490 × 65
828
Weight 441g 900g 1000g 36kg 12kg
Important criteria for the division is also the way how you communicate with
the device. The most comfortable are the ones that use wireless connection - uti-
lizing Bluetooth technology. Below there’s a brief description of functions of body
composition analyzers.
Important criteria for the division is also the way how you communicate with the
device. The most comfortable are the ones that use wireless connection - utilizing
Bluetooth technology.
Below there’s a brief description of functions of body composition analyzers.
For the parameters described in the above table, the algorithms used to deter-
mine parameters and their verification at the design level with the so-called Gold
Standard are very important, that is in our case, comparison with DXA binary x-
ray absorption. Much more useful from user service and analysis perspective are
those that allow the measurement of a large amount of data. For those who train and
care for their health, the muscle composition parameters are very important, espe-
cially after the data is divided and presented for every limb separately. This allows
for an optimal selection of training programme and training biasing to improve the
weaker muscle group. When used in cardiology, neurology, or in patients requiring
assessment of edema, it is helpful to measure a specific segment of the body.
Fig. 8.2: Examples of stationary units for measuring body composition[8][9]
Fig. 8.3: Examples of mobile analyzers of body composition [10].

Table 8.2: Comparison of devices for bioimpedance measurement.
Bodystat 1500 Akern BIA 101 IMPEDIMED Seca mBCA515 Tanita MC780
Anniversary SFB7
sport edition
TBW
Total body water
ECW
Entracellular water
ICW
Intracellular water
FM
Fat mass
FFM
Fat free mass
BCM × ×
Body cell mass
Muscle mass × × × × ×
BMI
FM
Fat mass
Bone mass × ×
BMR ×
Basal metabolic rate
Metabolic age × × ×
Skeletal muscle mass × × ×
BCMI × × × ×
8.6 Different methods for the measurement and analysis of body

composition
8.6.1 DXA
Imaging method utilizing three different techniques - double Xray beam (dual beam
- simultaneous high- and low-energy beam measurement), ultrasonography and
computer tomography. This method allows for accurate assessment of health sta-
tus, body composition and is the basis for the diagnosis of osteoporosis, osteopenia
and calcifications. This device is very accurate regarding the measurement such as
composition of soft tissue and bones, including bone mineral density, fat free mass,
fat mass and percentage of fat. Examinations that are carried out with used this
method are cheap, very precise and, in addition, the results have a high repeatabil-
ity rate. Densitometry was developed and refined by the FDA in the 1980s, and in
the span of 30 years there have been innovations such as narrow beam angles with
MVIR reconstruction - which eliminate the enlargement error and the need to use
direct to digital detectors to improve image quality. It also lowers the dose of ra-
diation when compared to competing systems. Those devices are characterized by
automated scanning and analysis functions and a full suite of powerful data man-
agement tools. Those include facilities such as multi-user database, batch output
(sending multiple tests to a DICOM printer), database export (export data to files
supported by Excel, for example [11]) . The radiation dose is very small, only 2.0
uSv. This dose is received by any of us just being for four hours outdoors. Neverthe-
less, pregnancy is a contraindication to this type of examination. Reproducibility of
measurements is only obtained on the same manufacturer’s equipment as they differ
in the evaluation algorithms used.
8.6.2 Antropometric measurements
They rely on the assessment of growth to weight ratio and the different skin fold
measurements. This method despite many drawbacks has many supporters It is
based on the premise that under the skin there is a constant fraction of body fat,
which is problematic especially in the elderly. The mass of fat tissue is assessed on
the basis of suitable arrays that were created on the basis of regression equations
and are specific for the population. Repeatability of this method is low.
TBW measurement is better at detecting fluctuations in extracellular fluid but
does not take into account changes in the intracellular potassium concentration as-
sociated with certain diseases. The accuracy of the data depends on the dose of
radiation.
8.6.3 Analysis of neuronal activation
Calcium and nitrogen isotopes that are captured by the nuclei of the body are used
for these studies.This allows us to determine calcium, sodium, chlorine, phosphorus,
and nitrogen content throughout the body.
8.7 Conclusion
Today’s design solutions have great potential when it comes to bioimpedance re-
search. The huge competition in the market for this type of equipment results in
new solutions, with better and better parameters that allow for more accurate mea-
surements, and thus improve the diagnosis of cardiovascular disease, kidney dis-
ease, underweight, overweight, as well as progress athletes or physically active per-
sons. In the Institute of Medical Technology and Apparatus, a device for measur-
ing bioimpedance spectroscopy has also been developed, which, based on a wide
range of impedance measurements at different frequencies, allows to map the entire
Cole Cole curve and measure body composition. Currently, further development
work is being carried out to improve the features of the device and optimize the
bioimpedance data evaluation algorithms.
References
1. Wayer, S., Zink, M., Wertzek, T.,and all: Bioelectrical impedance spectroscopy as a fluid man-
agement system in heart failure. ,IOP science, 817–930 (2014)
2. Kaysen, G., Zhu, F., Sarkar, S., and all: Estimation of total-body and limb muscle mass in
hemodialysis patients by using multifrequency bioimpedance spectroscopy.
The american journal of clinical nutrition,Warszawa, (2017)
3. Somma, S., Lukaski, H., Codognotto, M., and all: Consensus paper on the use of BIVA in
medicine for the management of body hydration, , energency care journal, 6–14 (2011)
4. http://www.rcz-zbaszyn.pl/dla_pacjentow/aktualnosci/191/
podstawy_teoretyczne_analizy_impedancji_bioelektrycznej.html
5. Khalil, S., Mohktar, M., Ibrahim, F.,: The theory and Fundamentals of bioimpedance analysis
in clinical status monitoring and diagnosis of dieses. Sensors 10895–10928 (2014)
6. Szulc, A.,: Biophysical and environmental basis of the analysis of body composition . Journal
of Health Sciences 410–427 (2013)
7. Bioelectrical impedance measurement in patients with gastrointestinal disease: validation of the
spectrum approach and a comparison of different meted for screening for nutritional
8. https://www.seca.com/pl_pl/produkty/wszystkie-produkty/
szczegoly-produktu/secambca515.html
9. http://analizatorybia.pl/tanita/295-tanita-mc-780-ma.html
10. http://analizatorybia.pl/impedimed/293-impedimed-sfb7.html
11. http://www3.gehealthcare.pl/pl-pl/produkty/kategorie/stan_
kosci/dxa/prodigy_for_bone_health
12. http://www.kcm.pl/wp-content/uploads/2016/01/A.6.1_
Densytometria-pacjent-29-PORADNIA-2015.11.07ec.pdf
Chapter 9
Prototype measurement system for spatial
analysis of speech signal for speech therapy
Kinga Kostera, Wojciech Wi˛eclawek, and Michał Kr˛ecichwost
Abstract. The paper provides a description of a measuring system for multi-channel

speech signal analysis. The aim of the work was to construct a spatial recording
station and prepare a processing software.
A developed measurement system consists of multi-channel recording compo-
nents with a microphone array, a connection to a computer, and a development com-
ponent which implements Delay-Sum Beamforming technology. Spatial-temporal
processing of the signal, i.a., determined the delay between the microphones by
calculating the offset for which the cross-correlation function between the signals
reaches the maximum.
The workstation was tested in order to verify the correctness of the system. The
testing procedure included recording at different distances (5 cm and 10 cm) and for
two voices (male and female). The results are satisfactory and confirm the proper
operation of the system. The best results were achieved for recordings of woman’s
voice at a distance of 5 cm from the sound source. This suggests that the measure-
ment system could be employed for the speech test of children whose voice tract is
shorter and speech is higher compared to male speech.
9.1 Introduction
Microphone arrays concern the application of many sensors to transmit or receive

a signal transferred by propagating waves. The application of microphone arrays is
necessary to create a measurement system for analysis or sound monitoring. The
result of using a microphone array is the multi-channel signal [1].
Methods based on multi-channel acoustic measurements are used in a variety of
industrial processes like diagnostic systems for energy transporters, environmental
noise monitoring, recording and amplification of the signal in hearing aids [2]. The
Faculty of Biomedical Engineering, Silesian University of Technology, Roosevelta 40, 41-800

Zabrze, Poland

https://doi.org/10.1007/978-3-319-70063-2_9
80 Kinga Kostera et al.
acoustic signal is an important source of information on many planes and that is

why this subject is so common [3].
Microphone arrays are also beginning to be used in speech therapy. Speech, or
acoustic signal, is the main diagnostic source [4]. The health assessment is not mea-
sured or objectified. With the measurement system in the form of a microphone
array it is probably possible to diagnose the most common speech defects. Proper
diagnosis of pronunciation defects is crucial to the health of a patient. Speech de-
fects have a negative impact, i.a., on the process of writing and reading [5].
The aim of this study is to design a prototype measurement system to spatial
analysis of speech signal for speech therapy. Our main contribution is building a
linear microphone array for recording speech signals. The numerical signal anal-
ysis will be performed to record signals from the multi-channel system, apply the
selected acoustic beam shaping algorithm, and visualize the results.
9.2.1 Database
It was necessary to gather the relevant database. The authors collected the speech
record database of two speakers – woman and man. The sequence ’ASA ESE ISI’
was proposed by speech therapists. The recordings were made from a distance of 5
cm and 10 cm. The sampling rate of the received signals was 44100 Hz.
9.2.2 Measuring system
A simplified scheme of the measuring system consists of a data acquisition system

and computer with software and is presented in Fig. 9.1.
Electromagnetic microphones POM-3044P-R do not require high polarization
voltage and are characterized by low efficiency. The frequency range of selected
microphones covers almost the whole audio spectrum (20 Hz - 20 kHz). Micro-
phones are characterized by a sensitivity -44 dB and have very small dimensions
(6 x 3 mm). In addition, the input impedance is 2.2 kΩ , working at a maximum
input voltage of 2 V. A multi-channel amplifier based on seven integrated circuits
NE5532D was designed. The system used consists of two low-noise operational
amplifiers. It allows for a two-stage gain (40x each) on a single audio channel. The
digital signal is obtained through a 16-bit USB-1608FS A/D converter.
The microphone array design is presented in Fig. 9.2. The linear array of micro-
phones is created by a set of 7 microphones arranged in the line at equal intervals.
The first array sensor (mic. 1) is the reference microphone. The source of the
sound is always opposite to it. The distance between the microphones was deter-
9 Prototype measurement system 81
Fig. 9.1: Measuring station
Fig. 9.2: Microphone array
mined on the basis of the pattern of the delay, which after the transformations looks
as follows [7]:
|r − li |
Δi = (9.1)
v

number, i = 1, ..., 7,, li – distance from sound source to i mi-
where i – microphone th
crophone, li = r2 + xi2 , r – distance from sound source to reference microphone,

v – sound velocity in air (340 m/s), xi – distance from ith microphone to reference
microphone.
9.2.3 Signal processing
Creating a system to support speech therapy requires a multi-step methodology.

Pre-processing of the signal recorded through the multi-channel system will be con-
sidered first. In Fig. 9.3 the general block diagram of the established methodology
is presented.
Fig. 9.3: Block diagram
The rst stage of signal processing was ltration. Signal was ltered using a lter
with a nite impulse response FIR. The cut-off frequency of lter is 400 Hz. Filtra-
tion is also used to cut-off specic frequencies, such as high noise or interference.
Delay-Sum Beamforming (DSB) is a classic algorithm for beam steering. This
method involves adding a delay associated with the change of the sound source
relative to the reference direction [8]. Calculating the delay is equivalent to setting
the cross-correlation function. Correlation is a measure of similarity between two
signals: +∞
R(τ ) = x(t)y(t + τ )dt (9.2)
−∞
where: x i y – comparable signals, τ – time delay.
The maximum value of autocorrelation occurs at the time delay. The autocorre-
lation function can be used to accurately measure the delay [9].
9.3 Experiments and results
A series of tests has been completed. The rst test was a recording from 5 cm and
10 cm. Then a test was performed using signals of different frequencies. This simu-
lation was executed with use of male and female speech.
9.3.1 Distance from sound source to microphone array
Results of using the cross-correlation function between the signals from each chan-
nel and the signal from the rst (reference) microphone are presented in Tab. 9.1.
Table 9.1: Effect of matrix distance from source to delay and amplitude of signals
from the following channels
Distance [cm] 5 10
Channel No 1 2 3 4 5 6 7 1 2 3 4 5 6 7
Delay [samples] 0 -2 -7 -12 -17 -24 -30 0 -2 -6 -10 -16 -22 -28
Amplitude 22.49 16.77 7.81 5.82 4.54 2.10 1.21 8.55 7.39 4.38 3.46 2.93 1.40 0.81
Measurement of the correctness of the time offset compensation is an increase

in the amplitude of the summed signal. In Figs. 9.4 and 9.5 the summation result
before and after time offset compensation is presented.
Fig. 9.4: Summed signal before and after time offset compensation for real speech
signal from distance equal 5 cm
Fig. 9.5: Summed signal before and after time offset compensation for real speech
signal from distance equal 10 cm
9.3.2 Influence of recorded signals frequency
The use of the cross-correlation function with the recorded signals and the signal
from the first microphone for both male and female speech is presented in Tab. 9.2.
The sum of the signals from the individual channels before and after the time delay
compensation is presented in Figs. 9.6 and 9.7.
Table 9.2: Effect of signal frequency to delay and amplitude of signals from the
following channels
Pitch Female speech (upper tone) Male speech (lower tone)

Channel No 1 2 3 4 5 6 7 1 2 3 4 5 6 7
Delay [samples] 0 -1 -3 -5 -8 -11 -14 0 -1 -3 -5 -8 -11 -14
Amplitude 1.92 1.60 1.03 0.85 0.76 0.34 0.31 2.94 2.73 1.74 1.40 1.20 0.60 0.47
Fig. 9.6: Summed signal before and after time compensation for the female voice
signal
9.4 Discussion
The signal amplitude is getting smaller in the following channels (Tabs. 9.1 and 9.2).
The similarity between signals of neighbouring channels is visible. The delays be-
tween signals from individual channels are imperceptible small. Differences be-
Fig. 9.7: Summed signal before and after time compensation for the male voice
signal
tween adjacent channels of the same recording are higher for each successive chan-
nel, which is an evidence of the spherical character of the acoustic wave.
The amplitude of the signal recorded from a distance of 5 cm is higher than that
recorded at 10 cm (Tab. 9.1). This difference is diminutive and is insignificant for
further calculations, but it is proof of the accuracy of the measuring system. The ef-
fect of time compensation is consistent with the prediction. Both time-compensated
signals have been weighted, what is shown in Figs. 9.4 and 9.5.
Cross-correlation values for male and female speech are significantly different
in amplitude, with higher values for male speech, what is shown in Tab. 9.2. This
may indicate a stronger of the male voice or the fact that the person pronounced the
sequence louder. Time delays are identical for both registrations, which confirms
the fixed distance of the sound source from the microphone array. The summation
result in Fig. 9.6 and 9.7 confirms the difference in amplitude values of the recorded
signals. Furthermore, a noticeable increase in the amplitude of the summed signals
is yielded by a proper compensation of the individual channel signals.
9.5 Conclusion
The aim of the work, which was to design a prototype measurement system to spatial
analysis of speech signal for speech therapy, was achieved. A measurement system
was developed, consisting of a hardware component allowing multi-channel speech
recording and transmission to a PC, as well as a developmental component that
implements Delay-Sum Beamforming. The developed workstation was subjected to

many tests. They were implemented to verify the correctness of the system. The
results are satisfactory and confirm the correct operation of the system. The best
results were achieved for recording the female voice at a distance of 5 cm from the
sound source. This means that the measurement system would perform very well in
the speech test of children whose vocal tract is shorter.
Acknowledgements
The work has been partially financed by Polish Ministry of Science and Silesian
University of Technology statutory financial support for researchers BK 200/RIB1/2017.
References
1. Krol, D., Lorenc, A., Swiecinski, R. Detecting laterality and nasality in speech with the use of
a multi-channel recorder, In: Acoustics, Speech and Signal Processing (ICASSP), 2015 IEEE
International Conference on, 2015.
2. B. Widrow, F. Luo, Microphone arrays for hearing aids: An overview, Stanford University,
Stanford, USA, Quicksilver Technology, San Jose, USA, 2003.
3. F. J. Fahy, The vibro-acoustic reciprocity principle and applications to noise control, Acustica
81:544, 1995.
4. C. Valentini-Botinhao, S. Degenkolb-Weyers, A. Maier, E. Noeth, U. Eysholdt, T. Bocklet,
Automatic detection of sigmatism in children, 2012.
5. Z. A. Benselam, M. Guerti, M. Bencherif, Arabic speech pathology therapy computer-aided
system, Journal of Computer Science, vol. 3, no 9, 2007.
6. M. Krecichwost, Z. Miodonska, J. Trzaskalik, J. Pyttel, D. Spinczyk, Acoustic mask for air
flow distribution analysis in speech therapy, In: E. Pietka, P. Badura, J. Kawa, W. Wieclawek,
Information Technologies in Medicine, pp. 377–387, Silesian University of Technology, Fac-
ulty of Automatic Control, Electronics and Computer Science, Gliwice, Springer International
Publishing, 2016.
7. K. Bogdzinski, M. Jasinski, J. Maczak, Projekt nisko-kosztowej macierzy mikrofonow pomi-
arowych, Zeszyty naukowe instytutu pojazdow, Warszawa, 2014.
8. D. Krol, R. Wielgat, Enhancement of loudspeaker impulse response measurement using beam-
forming methods, In: International Conference on Signals and Electronic Systems, ICSES, 18–
21, 2012.
9. S. Wisdom, T. Powers, L .Atlas, J. Pitton, Enhancement and recognition of reverberant and
noisy speech by extending its coherence, Department of Electrical Engineering, University of
Washington, Applied Physics Laboratory, University of Washington, 2015.
Chapter 10
Shear Wave Elastography and Strain
Elastography: a Study on a Phantom
Bartłomiej Pyciński1 , Jan Juszczyk1 , and Rafał Obuchowicz2
Abstract. In this paper we present a preliminary comparative study of two types of

ultrasound elastography techniques – Strain Elastography and Shear Wave Elastog-
raphy. The latter is a novel method which create a quantitative 2D map of the tissue
stiffness. The map is presented in absolute units (kPa). We proposed the layer-build
tissue-mimicking phantom to ensure repeatability of tests for both devices. The ob-
tained results confirms advantages of Shear Wave Elastography over Strain Elastog-
raphy.
Keywords: Shear Wave Elastography, Strain Elastography, Ultrasonography, Phan-

tom study
10.1 Introduction
Current ultrasonography (US) gives us only limited diagnostic accuracy in the field
of differentiation of the soft tissue lesions. Elastography, due to assessment of tissue
stiffness increases diagnostic accuracy permitting the highest diagnostic confidence
[1].
Quasi-static (strain) elastography (SE, also called “classical elastography” or,
simply, “elastography”) shows relative stiffness of tissue superimposed on B-mode
image in a region of interest (ROI). Ultrasound strain elastography is based on de-
tection of differences between tissue shifting on images consecutive in time under
pushing by ultrasound probe.

· Department of Radiology, Collegium Medicum,
Jagiellonian University, Cracow, Poland
e-mail: bartlomiej.pycinski@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_10
88 Bartáomiej Pyciński et al.
Shear Wave Elastography (SWE) is a technique, where an acoustic pressure wave

induces slow-moving lateral waves within the tissue, and the propagation speed of
the shear wave is proportional to the square root of the tissue’s elastic modulus.
Therefore shear wave velocity is faster in stiffer tissue and smaller in softer tissue
[2]. SWE is thought to be more objective than the previous elastography techniques
[3].
Quantitative measurements of tissue stiffness in absolute units (kPa) possible due
to acoustic radiation force push technique is the most objective method of tissue
density analysis, also when compared with strain elastography.
SWE technique has been presented in medical diagnosis since the ’90s [4], but
it has developed intensively since last few years [5]. Several phantom studies have
been recently presented [6, 7].
In this study we present a preliminary results of SE and SWE comparision. We
performed statistical analysis of SWE images manually and with tools integrated
into US device. A colloid phantom was prepared and its images were analysed.
Two ultrasound machines were used for ultrasound images and elasticity maps
recording. For Strain (Quasi-static) Elastrography a Philips iU22 device with lin-
ear probe L17-5 was employed and for Shear Wave Elastography a Supersonic Aix-
plorer with convex probe XC6-1 was used. Total number of 16 images were acquired
by the former and 98 by the latter. The image recording procedure was the same for
both systems.
Ultrasound and elastography images were captured in various positions of the
probe relatively to the phantom:
• at top and bottom walls to show layers of phantom along image depth direction
(position A in Fig. 10.1)
• at side wall vertically to show layers of phantom as a stripes on US image
(position B in Fig. 10.1)
• at side wall horizontally to show layers of phantom separately (position C in
Fig. 10.1)
10.2.1 US phantom
The colloid phantom consists of gelatin and bread flour (Fig. 10.1). It contains four
layers (α , β , γ , δ ) of different stiffness. Proportions of gelatin, water and flour for
each layer are shown in Table 10.1. The β layer’s stiffness is the highest. A differ-
ence between stiffness of this layer and the others’ can be distinguished by palpa-
tion. Thickness of layers equals to 6, 9, 10 and 12 mm for α , β , γ , δ , respectively.
10 Shear Wave Elastography and Strain Elastography: a Study on a Phantom 89
Fig. 10.1: a) Experimental scheme; b) The phantom’s photography. The layers are
marked with Greek letters
Table 10.1: Composition of the phantom’s layers
Layer No. Water [g] Gelatine [g] Flour [g]

α 100 10 10
β 150 20 30
γ 100 10 20
δ 200 10 40
Physical properties of the phantom affect velocity of US wave propagation:

Me
c= , (10.1)
ρe
where Me denotes effective elastic modulus and ρe denotes effective mass den-
sity [8]. The velocity affects the depth-size of the image because of constant speed
assumption. For soft tissue the velocity value is 1540 ms . If the true velocity differs,
axial dimension of the image is not computed properly.
Although precise knowledge about true ultrasound velocity in the phantom is
not necessary for this study, these experiments has also been performed (Section
10.3.3).
10.3 Experiments and results
10.3.1 Shear Wave Elasticity Imaging
For SWE testing, Aixplorer ultrasound device (Supersonic Imagine) was employed.
The images were acquired with XC6-1 convex probe. Image depth was set to 5
cm and the pixel size was equal to (0, 123 × 0, 123) mm/pix. Elasticity window
was overlaid on B-mode image with zero-opacity and linear greyscale colormap
ranged from 0 kPa (black) to 130 kPa (white). Image analysis was performed
on DICOM files. For image decompression (JPEGLossless Transfer Syntax) we
used GDCM library (http://gdcm.sourceforge.net/). Pixel value was
the “Value” channel of RGB to HSV data conversion. For ROI detection Python3
with Pydicom and NumPy were employed. Statistical analysis was performed with
R software (https://www.r-project.org/).
10.3.1.1 SWE image repeatability
At first, test–retest reliability was checked. US probe was fixed in stable position at
top side of the phantom and 14 images were captured with frequency 0.25 Hz. Pixel
values of layers α and β inside rectangle region of interest (ROI) of size 40 × 40
pix (Figure 10.2) were acquired. The same experiment was repeated at bottom side
of the phantom. Pixel values of the ROI located in layers γ and δ were acquired.
Summary data are presented in Table 10.3.
Fig. 10.2: Rectangle ROI for image repeatability testing. Appropriate layers are de-
noted by the symbols.
Table 10.2: Quantitative statistics of the layer’s images: α (top-left), β (top-right), γ

(bottom-left), and δ (bottom-right). Extrema of each column were bolded: minimal
values of minimum, 1st quartile and median as well as maximal values of median,
3rd quartile and maximum
no. min Q25 med Q75 max no. min Q25 med Q75 max
1 108.6 114.2 115.7 116.7 118.8 1 99.9 100.4 102.5 104.0 107.1
2 105.5 113.7 115.7 117.3 119.3 2 92.8 95.3 96.9 97.9 100.4
3 109.1 114.2 115.7 117.3 118.8 3 91.3 95.8 97.9 99.4 104.5
4 109.6 115.7 116.7 117.8 118.8 4 93.3 97.4 99.4 102.0 110.6
5 105.5 111.1 113.2 115.3 118.8 5 88.7 92.8 94.8 97.4 102.0
6 106.5 112.2 114.2 116.7 119.3 6 94.8 96.9 98.4 102.0 105.5
7 106.5 114.2 117.3 117.8 119.3 7 90.7 97.4 99.9 103.5 109.1
8 108.1 114.2 115.7 116.7 118.8 8 96.4 98.4 100.9 103.5 108.1
9 105.0 112.7 115.7 117.3 118.8 9 92.8 96.4 97.9 99.9 104.0
10 105.5 111.6 114.2 116.2 118.8 10 98.9 99.9 99.9 100.4 110.1
11 106.0 110.6 113.9 116.2 118.8 11 93.3 98.4 99.9 101.5 107.1
12 100.4 109.1 113.2 115.7 117.3 12 98.4 100.4 101.5 103.5 107.6
13 102.0 109.1 112.7 114.2 116.2 13 100.4 102.0 103.5 107.6 113.2
14 106.5 111.6 113.2 113.7 114.7 14 94.3 96.4 97.9 98.9 102.5
For each variable in each tested layer (α , γ , δ ) the Lillie test was performed in
order to determine, whether the variables had followed a Gaussian distribution. In
all cases the distribution was not normal, so the Kruskall-Wallis was applied in each
set. In each layer there was a statistically significant difference between the medians
(p-value < 10−4 ). Box-plot for layer α is shown in Figure 10.3.
Significant difference between the ROI intensities results from large cardinalities
of the datasets. For layer α , minimal and maximal median values are equal to 112.7
and 117.3 kPa, respectively. With the differences smaller than 5%, for clinical point
of view, the images are repeatedly well enough. Similar results are obtained for other
layers.
10.3.1.2 Elasticity of phantom layers
Elasticity of the phantom layers was also acquired using a measurement tool in-
tegrated into the Aixplorer device – Q-BoxTM . It shows a moveable and resizable
circle (ROI) on the colour elasticity map. The circle is duplicated on the same re-
gion of interest on the B-mode image for reference purposes. This tool measures
maximum, minimum, mean and standard deviation elasticity values in the region of
interest [9].
The values were obtained from various US probe position – at top, bottom and
side walls of the phantoms. 15 ROI-s of α layer, 11 ROI-s of β layer, 11 ROI-s of γ
layer and 5 ROI-s of δ layer were taken (Figure 10.4).
In Table 10.4 aggregated values were presented. Average standard deviations
were calculated under assumption that the data were uncorrelated. The results
Table 10.3: Quantitative statistics of the layer’s images: α (top-left), β (top-right), γ

(bottom-left), and δ (bottom-right). Extrema of each column were bolded: minimal
values of minimum, 1st quartile and median as well as maximal values of median,
3rd quartile and maximum
1 108.6 114.2 115.7 116.7 118.8 1 99.9 100.4 102.5 104.0 107.1
2 105.5 113.7 115.7 117.3 119.3 2 92.8 95.3 96.9 97.9 100.4
3 109.1 114.2 115.7 117.3 118.8 3 91.3 95.8 97.9 99.4 104.5
4 109.6 115.7 116.7 117.8 118.8 4 93.3 97.4 99.4 102.0 110.6
5 105.5 111.1 113.2 115.3 118.8 5 88.7 92.8 94.8 97.4 102.0
6 106.5 112.2 114.2 116.7 119.3 6 94.8 96.9 98.4 102.0 105.5
7 106.5 114.2 117.3 117.8 119.3 7 90.7 97.4 99.9 103.5 109.1
8 108.1 114.2 115.7 116.7 118.8 8 96.4 98.4 100.9 103.5 108.1
9 105.0 112.7 115.7 117.3 118.8 9 92.8 96.4 97.9 99.9 104.0
10 105.5 111.6 114.2 116.2 118.8 10 98.9 99.9 99.9 100.4 110.1
11 106.0 110.6 113.9 116.2 118.8 11 93.3 98.4 99.9 101.5 107.1
12 100.4 109.1 113.2 115.7 117.3 12 98.4 100.4 101.5 103.5 107.6
13 102.0 109.1 112.7 114.2 116.2 13 100.4 102.0 103.5 107.6 113.2
14 106.5 111.6 113.2 113.7 114.7 14 94.3 96.4 97.9 98.9 102.5
1 70.9 73.4 73.9 74.4 76.0 1 22.9 30.6 31.6 32.6 33.6
2 72.9 74.4 74.4 74.9 78.0 2 21.9 30.6 31.6 32.6 34.7
3 70.4 73.4 73.9 74.4 76.0 3 23.5 30.1 32.6 33.1 34.7
4 71.9 74.4 74.9 77.0 79.0 4 21.4 30.1 31.6 32.6 34.7
5 68.8 72.4 73.4 73.4 74.9 5 22.4 30.1 31.6 32.6 34.7
6 65.8 70.9 73.9 74.9 76.0 6 24.0 30.1 30.1 30.6 33.1
7 68.3 72.4 74.4 74.4 76.0 7 22.9 30.1 30.6 31.6 33.6
8 65.3 71.4 73.9 74.4 75.5 8 24.0 30.1 30.6 31.6 34.7
9 64.7 69.3 71.9 73.4 74.4 9 21.4 29.1 30.6 31.6 34.7
10 66.3 72.9 74.9 75.5 77.5 10 20.9 30.1 32.6 33.1 35.7
11 66.8 73.9 75.5 77.0 78.5 11 22.4 30.6 34.2 34.2 35.7
12 64.2 72.4 74.9 77.0 79.0 12 21.9 30.1 33.1 33.1 35.2
13 65.8 73.4 75.5 77.0 78.5 13 24.0 30.6 32.6 33.1 34.7
14 65.3 70.4 73.4 74.4 75.5 14 24.0 30.6 32.6 33.1 35.2
proved clear differences between the layers, although extreme values were spread
widely.
10.3.2 Strain Elastography
For strain elastography experiments Philips iU22 machine was used. The images
were acquired with L17-5 linear probe. Pixel size was equal to (0, 085 × 0, 085)
mm/pix. Elasticity image was presented side-by-side with B-mode image. Zero-
opacity of elasticity was used and linear greyscale colormap denoted white for the
Fig. 10.3: Box-plots of median and quartiles of the phantom’s stiffness. All images
of β layer’s (left); All layers (right)
Fig. 10.4: Tissue stiffness measured by the tool integrated into the US machine. The
phantom’s layer are marked with the Greek letters.
hardest and black for the softest region. Image analysis was performed on raw DI-
COM files. The probe was located at top side for images of α and β layer, and at
bottom side for the other layers.
As the elastrography data in this case is not quantitative, but rather relative to
actual maximal value, we did not perform advanced statistics. Elasticity intensities
of rectangle 40 × 40 pix ROI located in the middle of the layers were compared
(Figure 10.5).
Table 10.4: Overall quantitative informations of Q-Box ROI-s. Numbers of probe

directions are marked as follows: ↓ denotes the probe at top side directing down; ↑
denotes the probe at bottom side pointing up; Finally, → denotes the probe at side
wall.
layer no. mean min max st.dev. depth diam. direc-

[kPa] [kPa] [kPa] [kPa] [cm] [cm] tions
α 15 78.87 56.80 127.20 2.34 1.55 8.11 6 ↓, 9 →
β 11 85.80 51.40 117.30 4.10 2.33 5.36 6 ↓, 5 ↑
γ 11 74.39 52.20 133.60 2.61 2.55 5.00 6 ↓, 5 ↑
δ 5 26.04 18.10 37.00 2.83 0.78 7.00 5↑
Significant difference between the folowing ROI-s were proven: α < β and γ >
δ . These results are consistent with SWE findings.
Fig. 10.5: Quasi-static elastography image with appropriate layers denoted by the
symbols and the ROIs
10.3.3 Phantom ultrasound velocity
True height of the phantom was equal to 37.8 mm. The height obtained by Aixlorer
L10-2 linear probe was equal to 35.25. Given the ultrasound velocity used by the
machine equal to 1540 ms , we obtained approximate ultrasound velocity inside the
phantom equal to 1651 ms .
10.4 Conclusion
This paper summarize a preliminary study of shear wave elastography. This is rel-
atively new ultrasound technique which requires further investigation. Albeit the
repeatability of the measurement is not perfect, for clinical point of view the infor-
mation about the tissue elasticity given by SWE significantly outperform classical
elastography.
Preliminary observation revealed that there is a correlation between the localisa-
tion of the ROI and the elasticity values. Further investigation will be conducted.
Future research will cover also additional phantom measurement and in vivo ex-
periments.
Acknowledgement
This research is founded by Silesian University of Technology, Faculty of Biomed-

ical Engineering grant No. BKM-507/RIB1/2017.
References
1. DeWall, R.J.: Ultrasound elastography: Principles, techniques, and clinical applications. Criti-
cal Reviews in Biomedical Engineering 41(1) (2013) 1–19
2. Cosgrove, D.O., , Berg, W.A., Doré, C.J., Skyba, D.M., Henry, J.P., Gay, J., Cohen-Bacrie, C.:
Shear wave elastography for breast masses is highly reproducible. European Radiology 22(5)
(dec 2011) 1023–1032
3. Szczepanek-Parulska, E., Woliński, K., Stangierski, A., Gurgul, E., Biczysko, M., Majewski, P.,
Rewaj-Łosyk, M., Ruchała, M.: Comparison of diagnostic value of conventional ultrasonogra-
phy and shear wave elastography in the prediction of thyroid lesions malignancy. PLoS ONE
8(11) (nov 2013) e81532
4. Sarvazyan, A.P., Rudenko, O.V., Swanson, S.D., Fowlkes, J., Emelianov, S.Y.: Shear wave
elasticity imaging: a new ultrasonic technology of medical diagnostics. Ultrasound in Medicine
& Biology 24(9) (dec 1998) 1419–1435
5. Tian, J., Liu, Q., Wang, X., Xing, P., Yang, Z., Wu, C.: Application of 3d and 2d quantitative
shear wave elastography (SWE) to differentiate between benign and malignant breast masses.
Scientific Reports 7 (jan 2017) 41216
6. Pellot-Barakat, C., Chami, L., Correas, J.M., Lefort, M., Lucidarme, O.: Does motion affect
liver stiffness estimates in shear wave elastography? phantom and clinical study. European
Journal of Radiology 85(9) (sep 2016) 1645–1650
7. Rosskopf, A.B., Bachmann, E., Snedeker, J.G., Pfirrmann, C.W.A., Buck, F.M.: Comparison of
shear wave velocity measurements assessed with two different ultrasound systems in an ex-vivo
tendon strain phantom. Skeletal Radiology 45(11) (sep 2016) 1541–1551
8. Laugier, P., Haïat, G.: Introduction to the physics of ultrasound. In: Bone Quantitative Ultra-
sound. Springer Netherlands (nov 2010) 29–45
9. National Institute for Health and Care Excellence (NICE): Aixplorer shearwave elastogra-
phy for ultrasound imaging and assessing suspicious breast lesions (2014) Medtech innovation
briefing 15, nice.org.uk/guidance/mib15.
Chapter 11
A simple setup for repeatability analysis of
a low-level laser therapy scanner
Dawid Kucharski and Jagoda Nowak
Abstract. We present a study of the Polaris 2 device repeatability for a low-level

laser therapy (LLLT). We use a simple optical setup to test the repeatability of the
laser beam speed and profile and provide a statistical data analysis. Using our simple
optomechanic setup we show that the speed of the probing laser beam repeatability
is less than 8 % and the beam diameter repeatability less than 2 %.
Keywords: Low level laser therapy, repeatability, diode lasers, light emission
11.1 Introduction
Low level laser therapy (LLLT) is a well known biostimulation technique used
mainly in medicine [2], [3], [4], [10] and physiotherapy [9], [7], [5]. It is based
on laser-tissue interaction. Depending on the laser beam parameters such as wave-
length, power etc., light can stimulate the healing processes or to be use as a tool for
surgery. The physical mechanism of these influences on a human body is still a re-
search task [6], [8]. At present, the LLLT is used as a statistically proven technique
in medicine. The laser source in a LLLT scanner should be well known because it
determines the physical parameters such as: speed, beam geometry repeatability and
more.
We present a simple setup to test the repeatability of the laser light emission from a
Polaris 2 scanner. These preliminary results can be used to improve the construction
of future LLLT scanners for better determined physical properties of the laser light
and in this way led to understanding of the laser-tissue interaction effects.

Poznan ul. Piotrowo 3, 60-965, Poland,
e-mail: dawid.kucharski@put.poznan.pl,
WWW home page: http://wbmiz.put.poznan.pl

https://doi.org/10.1007/978-3-319-70063-2_11
98 D. Kucharski, J. Nowak
Fig. 11.1: Scheme of the experimental setup
11.2 Experimental Setup
11.2.1 The Polaris 2 scanner
In fig. 1 the scheme of the measurement setup is shown.

For the LLLT repeatability investigations we used a Polaris 2 laser scanner (see
fig. 11.2). In tab. 11.1 the laser radiation parameters are given.
The Polaris 2 laser light scanner has options for the treatment of many diseases and
is constructed for easy control of the treatment procedure. The operation parameters
can be chosen from the internal software for precise disease treatment (see fig. 11.3).
The scanner operator can set all parameters needed for treatment.
We determined the repeatability of the speed and the laser beam profile for three
different treatment programs.
Table 11.1: Polaris 2 laser radiation parameters
λ [nm] P [mW]
606 50
808 400
11 Light emission repeatability 99
Fig. 11.2: The Polaris 2 laser scanner Fig. 11.3: The Polaris 2 screen: 1 -
for LLLT: 1 - scanning head, 2 - tri- treatment field shape, 2 - distance to
pod, 3 - keyboard for a control, 4 - the treatment place, 3 - treatment field
lockable wheels width, 4 - treatment field length, 5 -
treatment time
Fig. 11.4: View of the detection setup Fig. 11.5: Detection setup: 1-2 - pho-
todiodes (DET36A/M), 3 - CMOS
camera (DCC1545M)
11.2.2 Detection
To probe the speed and laser beam profile repeatability we constructed a simple
optomechanical setup. In fig. 11.4, 11.5, 11.6 and 11.7 the components of this setup
are shown. For the scanning speed repeatability the two photodiodes (1 and 2 in
Fig. 11.6: Detection setup with the Fig. 11.7: Detection setup: 1- beam-
height of the phootodiodes indicated splitter plate (EBS1), 2 - circular vari-
able fused silica neutral-density filter
(CND-2-50.0M)
fig. 11.5) were used connected to an oscilloscope. For the geometry laser beam
repeatability a CMOS camera was used (3 in fig. 11.5). In tab. 11.2 and tab. 11.3 the
data for these detection components are given.
To reduce the too high intensity of the laser light we used a beamsplitter (1 in fig.
11.7) and a natural-density filter (2 in fig. 11.7) to avoid sensor damage.
Table 11.2: Photodiodes data
Active Area 13 mm2 (3.6 × 3.6 mm)

Wavelength Range 350 − 1100 nm
Rise Time 14 ns
Noise-Equivalent Power (NEP) 1.6 × 10−14 W/Hz1/2
Dark Current 0.35 nA (max. 6 nA)
Table 11.3: CMOS camera data
Resolution 1280 × 1024 Pixels

Frame Rate 25 fps
Dimensions (H x W x D) 48.6 mm × 44 mm × 25.7 mm
Pixel Size 5.2 × 5.2 μ m
Fig. 11.8: The laser beam path in rect- Fig. 11.9: The laser beam path in el-
angular scanning mode: 1-2 - photo- lipsoidal scanning mode: 1-2 - photo-
diodes, A-C - way of the beam move- diodes, A-C - way of the beam move-
ment during the rectangular scanning. ment during the ellipsoidal scanning.
Arrow corresponds to the laser prob- Arrow corresponds to the laser prob-
ing direction ing direction
11.3 Results
We tested the repeatability of the laser speed and the laser beam geometry of the
Polaris 2 scanner for different treatment settings and with two different ways of
scanning (rectangular and ellipsoidal scanning) (see fig. 11.8 resp. 11.9). The re-
peatability of the laser probing speed was investigated using the photodiodes. The
repeatability of the beam size (diameter) was determined using CMOS camera.
11.3.1 The speed of the beam
The repeatability of the speed was determined in both rectangular and ellipsoidal
beam scanning mode for three treatment programs. For the rectangular scanning in
the speed investigations, the change of the beam position is a double width of the
rectangle area of the scanning. For the ellipsoidal is an ellipse circuit.
In fig. 11.10 and 11.11 the voltages in time graphs are shown. Using the time infor-
mation, the velocity of the light movement was determined using the distance for
rectangular resp. ellipsoidal trajectory (see tab. 11.4).
Fig. 11.10: The voltage changes in time registered by a single photodiode in the
rectangular scanning mode for three different treatments: 1 - surgical scars, 2 - chon-
dromalacia patellae, 3 - tennis elbow
Fig. 11.11: The voltage changes in time registered by a single photodiode in the
ellipsoidal scanning mode for three different treatments: 1 - surgical scars, 2 - chon-
dromalacia patellae, 3 - tennis elbow
Table 11.4: Average speed of the laser beam for three treatment programs, calculated
using 110 periods
V [m/s]
Treatment program
Rectangular mode Ellipsoidal mode
Surgical scars 1.39 ± 0.11 7.46 ± 0.52
Chondromalacia patellae 1.42 ± 0.05 10.1 ± 0.5
Tennis elbow 1.39 ± 0.06 8.0 ± 0.5
The error given is 3σ
Fig. 11.12: A sample image of the ellipsoidal probing laser beam. As an example
on the image, the red dotted line corresponds to the direction of the pixel intensity
distribution shown in fig. 11.13
11.3.2 Laser beam geometry
The beam diameter repeatability was determined using images captured by the
CMOS camera and the pixel intensity distribution analysis. For image processing
we used the ImageJ [1] software which is freely available on the internet. To calcu-
late the beam diameter d the pixel intensity distribution was used combined with the
pixel size of the camera sensor (see tab. 11.3). In fig. 11.12 a sample image of the
beam is shown. We repeated this analysis 500 times per scanning mode to estimate
the average of the beam diameter (d) (see tab. 11.5). The image processing covered
a following steps:
1. capturing the images and store on a separate PC;
2. the pixel intensity distribution performed using ImageJ, as shown in fig. 11.12;
3. find the diameter of the laser beam using Gaussian profile fitting (see fig. 11.13).
Table 11.5: Diameter of the probing beam d for two scanning modes
d [μ m]
Treatment program
Rectangular mode Ellipsoidal mode
924 ± 16 923.7 ± 7.1
Surgical scars
915 ± 12 920 ± 18
The error given is 3σ
Fig. 11.13: The pixel intensity distribution example with gaussian curve for the
probing beam diameter (d) analysis
11.4 Summary
We present a simple and cheap optomechanical setup to test the Polaris II scan-
ner repeatability. We measured the speed and the laser beam diameter. The LLLT
scanner proves to have a good repeatability for two scanning modes for a different
treatment programs. The optical setup can be used in further investigations in better
understanding how a LLLT influences laser-tissue interaction.
Acknowledgements
The authors wish to thank prof. Ewa Stachowska and dr Frans Meijer for valuable
advice when preparing this article.
A work was supported by grant 02/22/DSPB/1387.
References
1. ImageJ - Image Processing and Analysis in Java. https://imagej.nih.gov/ij/, ac-

cessed: 2017-05-12
2. Avci, P.: Low-level laser (light) therapy (LLLT) in skin: stimulating, healing, restoring. Semi-
nars in cutaneous medicine and surgery 32(1), 41–52 (Mar 2013)
3. Avci, P., Gupta, G.K., Clark, J., Wikonkal, N., Hamblin, M.R.: Low-level laser (light) therapy
(LLLT) for treatment of hair loss. Lasers Surg Med 46(2), 144–151 (Aug 2013)
4. van Coevorden, R.S.: The low level laser in palliative care. Advances in Palliative Medicine
(2009)
5. Dawood, M.S., Al-Salihi, A.R., Qasim, A.W.: Laser therapy of muscle injuries. Lasers Med
Sci 28(3), 735–742 (May 2013)
6. Hode, L., Tunér, J.: Wrong parameters can give just any results. Lasers Surg Med 38(4), 343–
343 (2006)
7. Huang, Z., Chen, J., Ma, J., Shen, B., Pei, F., Kraus, V.B.: Effectiveness of low-level laser
therapy in patients with knee osteoarthritis: a systematic review and meta-analysis. Osteoarthr.
Cartil. 23(9), 1437–1444 (Sep 2015)
8. Kim, W.S., Calderhead, R.G.: Is light-emitting diode phototherapy (LED-LLLT) really effec-
tive? LASER THERAPY 20(3), 205–215 (2011)
9. Pinheiro, A.L.B., Oliveira, M.G., Martins, P.P.M., Ramalho, L.M.P., Oliveira, M.A.M.d.,
Júnior, A.N., Nicolau, R.A.: Biomodulatory effects of LLLT on bone regeneration. Laser Ther-
apy 13(1), 73–79 (2000)
10. Posten, W., Wrone, D.A., Dover, J.S., Arndt, K.A., Silapunt, S., Alam, M.: Low-level laser
therapy for wound healing: mechanism and efficacy. Dermatol Surg 31(3), 334–340 (Mar
2005)
Chapter 12
The influence of music genres on human
emotionality
Monika N. Bugdol1 and Marcin D. Bugdol1 Tomasz Smreczak
Abstract.
In the paper results of a preliminary research are presented concerning the in-
fluence of various music types on the human biocybernetic response. 35 people
(24 women, 11 men) of different age and occupation underwent the examination.
Each of them listened to three music pieces (classical, relaxing and rock music, in
a previously defined order). The measured biomedical signal was the GSR. Women
exposed more varied changes than men in their emotional state when they were
stimulated with different music types. Young people reacted strongest on classical
music, whereas mature persons revealed highest GSR changes while listening to
rock music. The results for different occupation groups are also presented in the
article.
Keywords: music therapy, GSR, music stimulation
12.1 Introduction
The influence of music on the human behavior, well-being, mood and mental com-
fort has been analyzed and applied since the beginning of humanity. At the begin-
ning of the development of music therapy there were no specialized music pieces
or specially selected melodies of appropriate tone, but simple, well-known sounds,
extracted by means of primitive instruments were played.
The first music therapists were shamans. They used music in magical and reli-
gious rituals to drive out "evil spirits" from a sick body. Fever was discarded using
simple, slow and rhythmic patterns to lower the body temperature. In the case of
rheumatoid pains, patterns were faster and more complex [1].
Faculty of Biomedical Engineering, Department of Informatics and Medical Equipment, Silesian

University of Technology, Zabrze, Poland
e-mail: monika.bugdol@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_12
108 Monika N. Bugdol et al.
One of the most famous music therapists was Descartes. He claimed that the
high intervals between sounds could cause emotional stress, while period sounds
can calm and encourage contemplation [2].
The rapid development of music therapy occurred during the Second World War
when soldiers and civilians had mental problems. Music was then officially recog-
nized as a therapeutic agent. It become compulsory to hire musicians in military
hospitals. Soon musicians working at hospitals were taught medical and psycholog-
ical issues, which implied the development of the profession known today as music
therapist. After the Second World War in Europe the "Swedish School", parallel
to the "American School" was established, educating in the field of music ther-
apy. The "American School", also known as "music pharmacology", represents an
empirically-clinical approach, based on the description of the reactions to different
music styles. "Swedish School" is based on depth psychology, due to the character-
istics of music that penetrates the deepest layers of personality [3].
Poland also contributes to the development of music therapy. In 1958 A. Demi-
anowski documented the use of music for the treatment of neurological disorders.
He paid particular attention to selecting appropriate music to match the mood of
the patients, which was aimed at relieving emotional tension and inducing general
relaxation.
J. Aleksandrowicz and S. Cywnar combined music therapy with hypnosis to re-
duce the dose without reducing the effect of the drug. Music can be therefore con-
sidered as an element that supports the effect of pharmacological drugs [4].
Under the scientific guidance of prof. Mitas the effects of sound stimulation on
brain efficiency are currently investigated [5]. Also the fetus well-being, that lives
in the womb of a mother, who listens to music is the subject of interest and research
[6].
There are a number of ways to measure the biocybernetic response of the human
body to music stimulation [7]. The most popular are Galvanic Skin Response (GSR),
Electrocardiography (ECG) and Electroancephalography (EEG) [8].
Galvanic Skin Response is a description of the change in the electrical properties of

the skin under the influence of the psychological state and well-being of the subject.
During GSR testing, the skin resistance or its opposite – the conductivity of the skin
– is measured [9].
The signals were acquired using Biopac MP36 from Biopac Systems with
the SS3LA sensor. This sensor consists of two Ag-AgCl electrodes mounted in
polyurethane housings. Velcro fasteners’ task is to facilitate the attachment of elec-
trodes to the fingers 12.1.
The electrodes are 6mm in diameter and 1.6mm deep. They have been polarized
to reduce interference and improve signal recording.
12 The inuence of music genres 109
Fig. 12.1: SS3LA sensor
The examined group consisted of 35 people aged 20-75, including 24 students,

5 pensioners, 3 salespeople, 1 teacher, 1 office worker and 1 worker of medical
industry. The research sample included 24 women and 11 men.
The test subjects were informed, that they would hear a minute melody, then a
15 second pause, another minute of melody, 15 second pause and the last minute of
music. Table 12.1 lists the compositions listened to by the respondents.
10 different sets on music tracks were defined, each consisting of three different
types of music pieces. Each of the sets 1-5 was listened to by 4 persons and sets
6-10 by 3 persons.
Table 12.1: List of music pieces
Classic Relaxing Rock

Ludwig van Beethoven Musica de Relajacion Janes Addiction
"Moon Sonata" el Bosque "Irresistible-Force"
Ippolitov Ivanov Sound of the sea Papa roach
"March of Teh Sardar" "Last Resort"
Richard Wagner Pretty Whores
"Ride of the Valkyries" "Midnight ShowdownÃćÂĂÂİ
Ludwig van Beethoven
"Violin Romance"
Giuseppe Tartini
"DevilÃćÂĂÂŹs Trill"
12.3 Results
Fig. 12.2 presents the standard deviations of conductivity of people, who listened
to the same set of music pieces, here is no 1. In case of this set first classical, then
relaxing and finally rock music was played. As it can be noticed, each examined
subject responded distinctively to different kinds of music stimuli. Similar diver-
sity of biocybernetic responses to stimulation with different music types was also
observed for the remaining sets.
Fig. 12.2: Diversification of the response to music set no 1
The following three graphs (Fig. 12.3, 12.4, 12.5) show the standard deviations
of the conductivity for the type of music in relation to the age of the studied subjects.
Young people reacted the strongest to classical music and then to relaxation music.
In case of strong sounding music, a noticeable increase in conductivity was observed
for mature people.
Fig. 12.3: The reaction of the respondents to classical music
Figures 12.6, 12.7, 12.8 show average values of standard deviations of conduc-
tivity for all subjects and with respect to sex and occupation
Analyzing the study in terms of the sex (Fig. 12.7) it can be concluded, that
both women and men react to the sound type. The average value of the standard
Fig. 12.4: The reaction of the respondents to relaxing music
Fig. 12.5: The reaction of the respondents to rock music
Fig. 12.6: Average standard deviations of conductivity for particular music

Fig. 12.7: Average standard deviations of conductivity for particular music kinds
with respect to sex
deviations of conductivity for women was highest, while they were listening to the
rock songs, it was almost equal to 0,12 μ S. The minimum value of the average
standard deviation of conductivity was also achieved by women and it was only 0.3
μ S. This value was recorder for classical music. The amplitude for women was 0.1
μ S, while for men it was 0.017μ S, which indicates that women are more sensitive
for music gender and they have tendency to change their emotion under the influence
of external stimuli.
The comparison of the mean standard deviation of the conductance with the divi-
sion of occupational group suggests that almost in every group the strongest reaction
can be noticed for the rock music. The only exception was students whose reaction
for rock music was the weakest. This may indicate that they prefer such kind of
music. The largest change in conductivity was observed in sellers while they were
listening rock music which reflects their strong reaction to this music. It can by
clearly stated that it caused anger, but it can be said that this type of music cased
the greatest emotions. The lowest value of the average standard deviation coefficient
can be noticed for the examined teacher – it was only 0,0085 μ S, achieved when he
was enjoying relaxation music.
Table 12.2 lists the number of persons who responded most and least strongly to
each type of music
Among 35 examined people, seventeen had the smallest galvanic skin response
while listening to relaxing music - 48% of respondents revealed the highest degree
of relaxation and only seven persons (20 %) showed the strongest reaction to this
genre. The lowest standard deviation of the conductivity while listening to rock
music was achieved by only 8 patients, (23%), which is more two times smaller
than for relaxation music. More than half of the respondents (51%) had the highest
values of standard deviation of conductivity while listening to rock music.
Fig. 12.8: Average standard deviations of conductivity for particular music kinds
with respect to occupation
Table 12.2: Number of people with the strongest / weakest response to a given type
of music
Number of people whose Number of people whose standard

Music standard deviation of conductivity standard deviation of conductivity
was the lowest was the highest
for a given type of music for a given type of music
Classical 10 10
Relaxing 17 7
Rock 8 18
12.4 Discussion
After the experimental study, it can be confirmed that each person reacts specifically
and individually to external sound stimuli. The analysis of the results leads to the
conclusion, that for most of the subjects there was an increased activity of sweat
channels while listening to rock music, than when relaxation music was played.
Gender effects on the GSR ware noticeable - women had more diverse reactions to
different types of music, more susceptible to calming relaxation music and energiz-
ing rock music. There were differences in biofeedback in various age groups - their
probable cause is the fact, that musical preferences evolve with time.
Future research will include a significant extension of the examined group. It
is also planned to acquire a wider variety of biomedical signals. A system will be
elaborated, which will enable simultaneous multi-channel registration of biomedical
signals with a built-in stimuli functionality.
References
1. Sachs C., Muzyka w świecie starożytnym, PWN, Warszawa 1981.

2. Galińska E., Dzieje pogladów
˛ na lecznicze działanie muzyki, Archiwum Historii i Filozofii
Medycyny nr 3, 1987, p. 235-255.
3. Szulc W., Szabelska-Holeksa M., Kształcenie muzykoterapeutów w USA i niektórych krajach
europejskich, Zeszyty Naukowe Akademii Muzycznej we Wrocławiu nr 76, Wrocław 2000,
s. 27-40.
4. Janicki A., Aktualny stan muzykoterapii w Polsce i w innych krajach świata – poglady,
˛ trendy,
szkoły, organizacja, kształcenie muzykoterapeutów, Zeszyty Naukowe Akademii Muzycznej
we Wrocławiu nr 73, Wrocław 1998, s. 76-82.
5. Mitas A.W., Lipowicz A.M., Wijata A.M., Bieńkowska M.J. (2016) Impact of Music on the
Effectiveness of Performing Mathematical-Logical Tasks. In: Pi˛etka E., Badura P., Kawa J.,
Wieclawek W. (eds) Information Technologies in Medicine. Advances in Intelligent Systems
and Computing, vol 471. Springer, Cham
6. Bieńkowska M.J., Mitas A.W., Lipowicz A.M. (2016) Model of Attenuation of Sound Stimuli
in Prenatal Music Therapy. In: Pi˛etka E., Badura P., Kawa J., Wieclawek W. (eds) Information
Technologies in Medicine. Advances in Intelligent Systems and Computing, vol 471. Springer,
Cham
7. Skotnicka M., Mitas A.W. (2014) About the Measurement Methods in Music Therapy. In:
Pi˛etka E., Kawa J., Wieclawek W. (eds) Information Technologies in Biomedicine, Volume 4.
Advances in Intelligent Systems and Computing, vol 284. Springer, Cham
8. Bugdol M. D., Mitas A. W.: Multimodal biometric system combining ECG and sound signals.
Pattern Recognition Letters 38,107–122 (2014)
9. Frederic Martini, Edwin Bartholomew: Essentials of Anatomy & Physiology. San Francisco:
Benjamin Cummings, 2003, s. 267. ISBN 0-13-061567-6.
Part II
Signal analysis
Chapter 13
2017 Monitoring and Teletransmission of
Medical-Data in Heart Failure. First Report
Mateusz Tajstra1 , Piotr Rozentryt1 , Elzbieta Gadula-Gacek1 , Jacek Niedziela1 ,

Elżbieta Adamowicz-Czoch1 , Aneta Ociessa1 , Adam Gacek2 , Arkadiusz Gwóźdź3 ,
Marcin Wilczek3 , Aleksander Płaczek3 , Konrad Wojciechowski4 , Adam Sokal5 ,
Zbigniew Kalarus5 , Mariusz Gasior
˛ 1 , and Lech Poloński1
Abstract. Number of analyses taken collectively have provided evidence that RPM
using non-invasive approach reduce relative risk of all-cause mortality by 15 - 40%,
and the risk of HF-related hospitalizations by 14 - 36%. Improvement of hospi-
talization was greater in stable as compared to newly decompensated patients and
approached 30%, reduction of mortality was more pronounced in patients newly
(< 28 days) discharged from the hospital due to acute episode of HF and was esti-
mated at 38%. These techniques were also cost-effective. Optimal RPM techniques
in HF are still not established. The large diversity of HF phenotypes likely differing
with respect to their pathophysiology precludes one common solution to prevent
HF decompensation and hospitalizations. Rather, various individually tailored tech-
niques should be searched for. The implantable cardiac device companies have their
own databases and follow-up systems. Users of the platform are assigned to support
levels corresponding with their duties and knowledge. We determined a model of
cooperation between the medical personnel in the monitoring and care of patients.
The users have their own lists of cases to manage. The incidents or follow-up trans-
missions are the cases to manage. The frst-line/primary support level analyzes the
data and determines a course of action.
Keywords: cardiology telemedicine, ITIL
Third Department of Cardiology, SMDZ in Zabrze, Medical University of Silesia in Katowice,

Silesian Centre for Heart Disease, Zabrze, Poland
· Institute of Medical Technology and Equipment, Zabrze, Poland
· WASKO SA, Gliwice, Polande-mail: M.Wilczek@wasko.pl
· Polish Japanese Academy of Information Technology, Warszawa, Poland
· Department of Cardiology, Congenital Heart Diseases and Electrotherapy, Silesian Center for
Heart Diseases, Medical University of Silesia, Zabrze, Poland

https://doi.org/10.1007/978-3-319-70063-2_13
118 Marcin Wilczek et al.
13.1 Introduction
Heart failure (HF) has becoming an epidemic globally [1]. Modern therapy has
slightly reduced mortality in HF, but at the expense of more frequent hospitaliza-
tions [2]. Numerous techniques of remote patients monitoring (RPM), either using
invasive [3] or non-invasive approach [4, 5] have been developed and tested world-
wide as a tools to guide early pre-emptive therapy and avoid costly hospitalizations.
Number of analyses taken collectively have provided evidence that RPM using
non-invasive approach reduce relative risk of all-cause mortality by 15 - 40%, and
the risk of HF-related hospitalizations by 14 - 36%[6, 7]. Improvement of hospi-
talization was greater in stable as compared to newly decompensated patients and
approached 30%, reduction of mortality was more pronounced in patients newly (<
28 days) discharged from the hospital due to acute episode of HF and was estimated
at 38%. These techniques were also cost-effective.
Apart from non-invasive RPM systems, some patients take advantage of im-
plantable electronic devices. Among that class, there are some specifically ded-
icated to monitoring filling pressures [8], some others, such as implantable car-
dioverter / defibrillators, with or without resynchronization function (ICD/CRTD),
may serve multiply purposes [9]. Beyond their key therapeutic capabilities, these
devices are additionally equipped in a wide range of diagnostic features. This al-
lows not only for reliable evaluation of device basic functions, but also assessment
of many physiological parameters, including in particularly the presence of ven-
tricular and/or supraventricular arrhythmias i.e. asymptomatic atrial fibrillation. Re-
cently published meta-analysis have shown that RPM from ICD/CRT-Ds reduces
number of planned hospitalizations, decreased costs of medical care, but slightly in-
creases unplanned hospitalizations without affecting mortality [10]. The RPM using
ICD/CRT-Ds have significant shortcoming. The indication for implantation is lim-
ited to small subgroup of patients with HF [11]. The can’t be used in patients with
preserved ejection fraction who significantly contribute to hospitalization burden.
Another drawback, frequently pointed out by experts, is the fact that ICD/CRT-D are
produced by different manufacturers, what prevents collection of signals (informa-
tion) sent from different devices in a single integrated database, thus limiting further
comprehensive standardized analyses [12]. These devices are being implanted ac-
cording to current guidelines in HF patients with appropriate indications irrespective
of their phenotypic differences, what in certain circumstances might lead to ambigu-
ous or even negative study results [13]. Optimal RPM techniques in HF are still not
established. The large diversity of HF phenotypes likely differing with respect to
their pathophysiology precludes one common solution to prevent HF decompensa-
tion and hospitalizations. Rather, various individually tailored techniques should be
searched for [14].
This idea has given the basis for initiation and development of “Monitoring and
teletransmission of medical data in patients with heart failure” program (MONITEL-
HF).
In short, the MONITEL-HF program contains developmental and clinical com-
ponents. The technical part was devoted firstly for creation of information platform
13 2017 Monitoring and Teletransmission. . . 119
capable of collection, transmission and storage of medical signals derived from HF

patients either from implanted devices or from non-invasive biosensors embedded
into wearable structure. Second technical part comprised of development of non-
invasive biosensors that can be integrated into textile and tested in terms of their
agreement with “gold standard” measurements and reproducibility and stability of
transmission. This task has been performed within “External tracking of clinical
instability using non-invasive biosensors and data teletransmission in patients with
heart failure”- EXTRABIOTEL-HF study. The clinical component of MONITEL-
HF comprised of prospective clinical, prospective randomized study of 600 HF pa-
tients with newly implanted ICD/CRT device called “Remote Supervision to De-
crease Hospitalization Rate” — RESULT study. The rationale and design of the
program has been extensively described in separate papers [15, 16]. We believe that
successful realization of key MONITEL-HF aims will pave the way for improve-
ment of patients’ quality of life, increase of their survival and allow reduction of
treatment costs, thus provide substantial savings in the national budget [17].
13.2.1 Information Platform for Integrated Data Collection
The platform was created by Wasko S.A. (Poland) in collaboration with physicians
from the Silesian Center for Heart Diseases. The implantable cardiac device com-
panies have their own databases and follow-up systems. They are implementing the
Integrating the Healthcare Enterprise Implantable Device Cardiac Observation (IHE
IDCO) profile and the IEE 11073-10103 nomenclature standard, which supports ter-
minology. The IHE IDCO profile has a standardized workflow.
Most of these companies implement communication with a local electronic
health records system (EHR)/EMR Hospital Management Systems using the Health
Level Seven (HL7) protocol. Wasko S.A. became a member of the HL7 Interna-
tional organization to understand and implement the standard in its own systems.
The platform architecture is based on ITIL methodology. In ITIL - level manage-
ment presupposes different time to solve the different kinds of incidents. The plat-
form is organized to support the different levels of support, including a primary
support level, secondary support level and tertiary support level.
Users of the platform are assigned to support levels corresponding with their
duties and knowledge. We determined a model of cooperation between the medical
personnel in the monitoring and care of patients. The users have their own lists of
cases to manage. The incidents or follow-up transmissions are the cases to manage.
The frst-line/primary support level analyzes the data and determines a course of
action. There are also screens that allow a physician to analyze a selected patient’s
transmissions, despite the workflow, to monitor a patient’s data. Producers name the
parameters and their values using the IEEE11073-10103 standard. In the reports,
there are parameters and values that enhance the IEEE standards and specific data
defined by the device producers. Some of the data, e.g., the graphic charts, is stored
in the embedded binary data as pdf files.
13.2.2 EXTRABIOTEL-HF study
By the day of paper preparation within EXTRABIOTEL-HF study several experi-

mental modules were developed by ITAM - technical research partner of consortium
to measure physiologic signals. The signals currently studied included:
• pulse wave velocity based on a new principle and parallel measurements of
peripheral blood oxygen saturation and ECG,
• blood oxygen saturation,
• stoke volume and cardiac output based on impedance reography,
• water content within an abdomen and chest cavity based on bioimpedance,
• heart rate and heart rate variability based on ECG,
• respiration rate,
• body position based on 3-axis digital accelerometer,
• galvanic skin response,
• skin temperature.
Testing measurements were performed in a group of 40 healthy subject and 50
persons with heart failure. Each biological signal of interest was obtained in parallel
using “gold standard” device and with newly developed modules.
Currently we perform the analyses of agreement between these pairs of signals
and assess their reproducibility.
The experimental modules developed so far are presented on Figure 13.1. These
modules are used for testing, the final product are intended to be highly miniaturised.
13.3 Results
The recruitment to the study was closed in October 2016. 600 patients with HF
were randomized into two groups: telemonitoring and of classic control. The base-
line clinical characteristics of patients randomized into study groups are shown in
Table 13.1.
13.4 Conclusions
Preliminary results show that home monitoring of patients with implanted high-
energy devices is safe and feasible in patients treated with ICD and CRT.
Fig. 13.1: Experimental modules developed to test selected biological signals.
Fig. 13.2: The impedance recording from testing module.

Table 13.1: Baseline clinical characteristics of the study groups.
Variable Remote monitoring p value

Yes No
(n=299) (n=301)
Age, years (IQR) 64.0 (13.0) 64.0 0.49
Male, n (%) 244 (81.6%) 243 (80.7%) 0.78
NYHA, n (%) 0.12
I 26 (8.8%) 24 (8.2%)
II 217 (73.3%) 198 (67.6%)
III 53 (17.9%) 68 (23.2%)
IV 0 3 (1.0%)
Device type, n (%) 0.89
Single-chamber ICD 91 (30.4%) 95 (8.2%)
Dual-chamber ICD 81 (27.1%) 84 (67.6%)
CRT-D 127 (42.5%) 122 (23.2%)
Device company, n (%) 132 (45.1%) <0.001
St Jude 133 (44.9%) 84 (28.7%) 1
Medtronic 118 (39.9%) 43 (14,7%)
Biotronik 37 (12.5%) 34 (11.6%)
Boston 8 (2.7%)
Comorbidities
DM, n (%) 108 (36.1%) 101 (33.6%) 0.51
Chronic kidney disease (GFR 72 (24.1%) 62 (20.6%) 0.31
< 60 ml/min/1.7 m2 ) (%)
History of AF, n (%) 99 () 92 (30.6%) 0.50
Hypertension, n (%) 187 (62.5%) 165 (54.9%) 0.0547
Ischemic etiology, n (%) 189 (63.2%) 195 (64.8%) 0.69
Non-ischemic etiology, n (%) 110 (36.8%) 106 (35.2%) 0.69
Stroke, n (%) 16 (5.4%) 12 (4.0%) 0.43
Prior MI, n (%) 146 (48.8%) 142 (47.2%) 0.69
Prior PCI, n (%) 166 (55.5%) 157 (52.2%) 0.41
Prior CABG, n (%) 61 (20.4%) 60 (20%) 0.89
Secondary prevention, n (%) 31 (10.4%) 20 (6.6%) 0.14
QRS, ms (IQR) 126 (53) 124 (42) 0.70
AF — atrial fibrillation; CABG — coronary artery bypass grafting; CRT-D - cardiac resynchro-
nization therapy-defibrillator; DM — diabetes mellitus; GFR - glomerular filtration rate; ICD —
implantable cardioverter-defibrillator; IQR — interquartile range; MI — myocardial infarction;
PCI — percutaneous coronary intervention
Acknowledgements
The MONITEL-HF Consortium under the leadership of Silesian Centre for Heart
Disease has received funding from National Center for Research & Development
under the frame of STRATEGMED project. The grant number: STRATEGMED1/233221/3/NCBR/2014
acknowledged on 20 October 2014.
Table 13.2: Analysis of end-points in both groups (mean observation time: 15

months).
Variable Remote monitoring p

Yes No
Composite end-point, n (%) 131 (43.8) 147 (48.8) 0.24
Death, n (%) 28 (9.4) 24 (8.0) 0.64
Hospitalization due to car- 122 (40.8) 139 (43.2) 0.21
diovascular reasons, n (%)
Amount of in-person visits in 4.4 5.7 <0.001
clinic (mean for each patient)
Cost of in-person visits 408.1 545.4 <0.001
in clinic (mean cost in
PLN/patient)
Cost of hospitalization due 26,951 21,921 0.04
to cardiovascular reasons,
(mean cost in PLN/patient)
Death and/or hospitalization due to cardiovascular reasons.
References
1. Ponikowski, P., Anker, S.D., AlHabib, K.F., Cowie, M.R., Force, T.L.: Heart failure: preventing
disease and death worldwide. ESC Heart Failure, 1, 4-25 (2014)
2. Fang, J., Mensah, G.A., Croft, J.B., Keenan, N.L.: Heart failure-related hospitalization in the
U.S., 1979 to 2004. J Am Coll Cardiol, 52, 428-34 (2008)
3. Adamson, P.B., Abraham, W.T, Stevenson, L.W., Desai, A.S., Lindenfeld, J.: Pulmonary Artery
Pressure-Guided Heart Failure Management Reduces 30-Day Readmissions. Circ Heart Fail, 9,
(2016)
4. Gastelurrutia, P., Cuba-Gyllensten, I., Lupon, J., Zamora, E., Llibre, C.:Wearable vest for pul-
monary congestion tracking and prognosis in heart failure: A pilot study. Int J Cardiol, 215,
77-9 (2016)
5. Shochat, M.K., Shotan, A., Blondheim, D.S., Kazatsker, M., Dahan, I.: Non-Invasive Lung
IMPEDANCE-Guided Preemptive Treatment in Chronic Heart Failure Patients: A Randomized
Controlled Trial (IMPEDANCE-HF Trial). J Card Failet, (2016)
6. Kitsiou, S., Pare, G., Jaana, M.: Effects of home telemonitoring interventions on patients with
chronic heart failure: an overview of systematic reviews. J Med Internet Res, 17, 63 (2015)
7. Conway, A., Inglis, S.C., Clark, R.A.: Effective technologies for noninvasive remote monitoring
in heart failure. Telemed J E Health, 20, 531-8 (2014)
8. Abraham, W.T., Stevenson, L.W., Bourge, R.C., Lindenfeld, J.A., Bauman, J.G.: Sustained e-
cacy of pulmonary artery pressure to guide adjustment of chronic heart failure therapy: com-
plete follow-up results from the CHAMPION randomized trial. Lancet 387, 45361 (2016)
9. Desai, A.S.: Implantable Hemodynamic Monitoring in Ambulatory Heart Failure: Who, When,
Why, How?. Curr Cardiol Rep 17, 113 (2015)
10. Klersy, C., Boriani, G., De Silvestri, A., Mairesse, G.H., Braunschweig, F.: Effect of telemon-
itoring of cardiac implantable electronic devices on healthcare utilization: a meta-analysis of
randomized controlled trials in patients with heart failure. Eur J Heart Fail, 18, 195-204 (2016)
11. Ponikowski, P., Voors, A.A., Anker, S.D., Bueno, H., Cleland, J.G.: 2016 ESC Guidelines for
the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diag-
nosis and treatment of acute and chronic heart failure of the European Society of Cardiology
(ESC)Developed with the special contribution of the Heart Failure Association (HFA) of the
ESC. Eur Heart J, 128, (2016)
12. Dubner, S., Auricchio, A., Steinberg, J.S., Vardas, P., Stone, P.: ISHNE/EHRA expert consen-
sus on remote monitoring of cardiovascular implantable electronic devices (CIEDs). Europace,
14, 278-93 (2012)
13. Van, Veldhuisen, D.J., Braunschweig, F., Conraads, V., Ford, I., Cowie, M.R.: Intrathoracic
impedance monitoring, audible patient alerts, and outcome in patients with heart failure. Circu-
lation, 124, 1719-26 (2011)
14. De Keulenaer, G.W., Brutsaert, D.L.: Systolic and diastolic heart failure are overlapping phe-
notypes within the heart failure spectrum. Circulation, 123, 1996-2004 (2011)
15. Rozentryt, P.G., Gwóźdź, A., Wilczek, A.Q., Chłosta-Niepiekło, M., Wojciechowski, K.,
Tajstra, K., Gadula-Gacek, M., Niedziela, E., Adamowicz-Czoch, J., Sokal, E., Ociessa, A.,
Kalarus, A., Gasior,
˛ Z., Poloński M.: MONItoring and TELetransmission of medical data in
Heart Failure. Rationale and design of the MONITEL-HF project. Interational Journal of Inter-
net Science, (2016)
16. Tajstra, M.S., Gwóźdź, A., Wilczek, A., Gacek, M., Wojciechowski, A., Gadula- Gacek, K.,
Adamowicz-Czoch, E., Chłosta-Niepiekło, E., Milewski, K., Rozentryt, K., Kalarus, P., Gasior,
˛
Z., Poloński L. M.: REmote SUpervision to Decrease HospitaLization RaTe. Unified and inte-
grated platform for data collected from devices manufactured by different companies: Design
and rationale of the RESULT study. Europace, (2016)
17. Gustafsson, F., Arnold, J.M.: Heart failure clinics and outpatient management: review of the
evidence and call for quality assurance. Eur Heart J, 25, 1596604 (2004)
Chapter 14
Application of Discrete Cosine Transform for
Pre-Filtering Signals in Electrogastrography
Dariusz Komorowski1 and Barbara Mika1
Abstract. Electrogastrography (EGG) is the technique of the cutaneous recording

of the myoelectrical activity of the stomach. Due to its noninvasiveness and correla-
tion with the gastric motility it is the attractive complement for imaging stomach’s
diagnostic methods. As the EGG signal is the mixture of the electrical activity of
the stomach and surrounding organs so the raw EGG also contains the noise, the
electrocardiographic (ECG), and the respiration (RESP) signals. The aim of this pa-
per is to present the effective tool for pre-filtering EGG signal. The filtering in the
Cosine Discrete Transform (DCT) domain has been proposed as an efficient tool
for denoising EGG signal. The obtained results are compared with the outcomes
determined by means of the traditional digital (Butterworth, in this case) filtering
method.
Keywords: EGG, Discrete Cosine Transform (DCT), filtering
14.1 Introduction
Electrogastrography is a method for recording the gastric myoelectrical activity us-

ing cutaneous electrodes placed on the anterior abdominal wall overlying the stom-
ach [1]. From the pacemaker area located on the grater curvature of the stomach
electrical excitation spreads circumferentially and distally down the stomach walls
[5]. The signal recorded by means of the electrogastrography technique is called
electrogastrogram (EGG). The EGG recording besides the leading rhythm, called
the normogastric rhythm (2.4–3.6 cpm), can also possess the some pathological
rhythms: bradygastric (<2.4 cpm), tachygastirc (3.6–9.0 cpm), and arrhythmia (the
ranges slightly differ, depending on the institution) [9]. Nowadays, the EGG signal
Silesian University of Technology, Faculty of Biomedical Engineering, Department of Biosen-

sors and Processing of Biomedical Signals, Zabrze, Poland, e-mail: dariusz.komorowski@
polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_14
126 Dariusz Komorowski, Barbara Mika
is a valuable complement to the other diagnostic methods of the stomach diseases

[2, 3, 11] especially those connected with the disturbance of the gastric motility.
The signals, which are available on the stomach surface include not only the EGG
signals but also the electrocardiographic signals, the signals connected with the res-
piratory movements, and the random noise [2]. The main component of the EGG
signal so called the slow wave has the frequency ∼3 cpm, that is, ∼0.05 Hz and
the amplitude in the range 50-500 μ V [6];thus, the EGG is the weak signal embed-
ded in the much stronger components such as ECG and RESP. In order to use EGG
as the diagnostic tool that interfering signals must be removed. Fig.18.1 shows the
frequency ranges of EGG, RESP, and ECG signals. As these ranges are partially
covered, there are some difficulties in choosing the effective method of processing
the EGG. Respiratory signal is considered to be the strongest disturbing factors in-
fluencing both the ECG and the EGG in the overlapped ranges of the frequency.
Typically EGG signal is extracted by means of the analog or the digital band-pass
Fig. 14.1: The frequency ranges of signals available on the stomach surface: EGG -
magenta, RESP - red , ECG - blue
filters. The range of frequency for EGG signal is from 0.015 Hz to 0.15 Hz (0.9-
9.0 cpm (cycle per minute)). Recently Hang Sik Shin [4] proposed Index-Blocked
Discrete Cosine Transform Filtering Method (IB-DCTFM) to design the ideal fre-
quency range filter in the DCT domain for the biomedical signals. IB-DCTFM re-
moves the unwanted frequency range signal in the time domain by blocking the
specific DCT index in the DCT domain. In this paper the IB-DCTFM algorithm is
suggested as the good performing method for attenuation the RESP signal from the
EGG data.
Fig.14.2 shows the exemplary 4-channel signal recorded from the surface of the
stomach, further called the raw EGG. It is easy to notice that the raw EGG contains
the ECG, RESP, and the random noise. The aim of this paper is to present the effec-
tive tool for pre-filtering EGG signal. The filtering in the Cosine Discrete Transform
(DCT) domain has been proposed as an alternative method to the commonly used
the filtering methods. The obtained results are compared with the results determined
by means of the Butterworth digital filtering [8].
14 Application of Discrete Cosine Transform. . . 127
4500
A1
A2
A3
4000 A4
3500
Recorded Signal (u.u.)
3000
2500
2000
1500
1000
00:00:00 00:00:01 00:00:02 00:00:03 00:00:04 00:00:05 00:00:06 00:00:07 00:00:08 00:00:09 00:00:10
time (hh:mm:ss)
Fig. 14.2: Exemplary of the recorded 4-channel signal from the surface of the stom-
ach (raw EGG)
14.2 Methods
The discrete cosine transform (DCT) express a the finite discrete sequence data of a
sum of cosine functions oscillating at different frequencies eq.14.1.
N
π (2n − 1)(k − 1)
y(k) = w(k) ∑ x(n)cos , k = 1, ..., N (14.1)
n=1 2N
where
1 2
w(1) = √ ∧ w(k) = f or 2≤k≤N
N N
N denotes the length of x and y, which are the same size. The filtering process
is conducted by applying the DCT transform and its inverse version IDCT. The
DCT provide the decomposition of EGG signal, so that the unwanted components
(frequencies) can be identified in the EGG signal and removed, whereas the IDCT
allows to reconstruct the EGG signal already without interferences eq.14.2.
N
π (2n − 1)(k − 1)
x(n) = ∑ w(k)y(k)cos 2N
, n = 1, ..., N (14.2)
n=1
where
1 2
w(1) = √ ∧ w(k) = f or 2≤k≤N
N N
The typical EGG examination takes above two hours and consist of three phases:
preprandial, meal, and postprandial. In order to assess the effectiveness of DCT
method the chosen parameter, such as, the normogastria index, obtained after DTC
filtering and traditional filtering have been compared. Due to the fact, that for the
mentioned parameter the calculation of the dominant frequency (DF) are required,
the power spectrum density (PSD) of EGG signal was determined after applying the
DCT and the traditional filtering method. The EGG signals have been filtered and
then split into 30-minute sections. Next, each fragment has been divided into 60 or
256-second segments and the power spectrum density (PSD) of the each segment
has been calculated and the dominant frequencies have been found [7]. In addition
the overall (average) PSD of the section has been calculated. Based on the overall
PSD, dominant frequencies for each section has been found and the normogastria
rhythm index (NI) defined, as the ratio of the number of segments for which DF is
in the range of 2.4 − 3.6
cpm (the range slightly differs, depending on the insti-
tution) [7, 9, 10] to the total number of segments has been calculated. The results
of the methods performance are presented for the exemplary EGG signal from the
postprandial phase after the stomach stimulation with the 400 ml of water.
14.3 Results
The methods was validated by the means of the real human signals. The results are
presented for the EGG signal recorded from the young 24 year old woman with the
body mass index (BIM) 19.9. The short fragments of EGG after the classical, in
this case, the Butterworth filtering and the DCT filtering are presented in the Fig.
14.3 and Fig. 14.4, respectively. Fig. 14.5 presented the examples of overall PSD
[7] of EGG signals without filtering, after using Butterworth filter (4th order, cutoff
frequencies 0.9–9.0 cpm) and the block DCT filter (cutoff frequencies 0.9–9.0 cpm).
250
A1
A2
200 A3
A4
150
100
Filtered EGGD Signal (uV)
50
−50
−100
−150
−200
00:00:00 00:00:20 00:00:40 00:01:00 00:01:20 00:01:40 00:02:00 00:02:20 00:02:40 00:03:00 00:03:20 00:03:40 00:04:00
time (hh:mm:ss)
Fig. 14.3: The 4-channel EGG signal after the Butterworth filtering
250
A1
A2
200 A3
A4
150
100
Filtered EGGD Signal (uV)
50
−50
−100
−150
−200
00:00:00 00:00:20 00:00:40 00:01:00 00:01:20 00:01:40 00:02:00 00:02:20 00:02:40 00:03:00 00:03:20 00:03:40 00:04:00
time (hh:mm:ss)
Fig. 14.4: The 4-channel EGG signal after the DCT filtering
Overall PSD(PER4) of EGG Signal, Channel (1)
3.018
−70
−80
Magnitude (dB)
−90
−100
−110
−120
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Overall PSD(PER4) of EGG Signal, Channel (1) Overall PSD(PER4) of EGG Signal, Channel (1)
3.018 3.020
−70 −70
−80 −80
Magnitude (dB)
Magnitude (dB)
−90 −90
−100 −100
−110 −110
−120 −120
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm) Frequency (cpm)
Fig. 14.5: The example of the overall PSD of the EGG signal (top-left) for origi-
nal EGG, after the Butterworth filtering (bottom-left), and after the DTC filtering
(bottom-right)
In order to verify the method performance the normogastria index (NI) [7] have
been calculated for the EGG signals processed by the traditional and the DCT meth-
ods. Fig. 14.6 and Fig. 14.7 present the distributions of the dominant frequency for
the 4-channel EGG signal processing by both the Butterworth and the DCT filtering.
In the Table 33.1 and 26.2 the values of NI index for all phases of the recorded EGG
signal, that is, preprandial, meal and postprandial obtained by the each of two kinds
of filtering are summarised. In addition, the postprandial phased has been divided
into the two 30-minute parts: postprandial 1, and postprandial 2, which corresponds
to the phase of the digestive process after the stomach stimulation with water (in
this case).
Max Power Spectrum, channel (1)

Magnitude (dB)
0
normo = 0.898
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Magnitude (dB)
0
normo = 0.816
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Magnitude (dB)
0
normo = 0.827
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Magnitude (dB)
0
normo = 0.969
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Fig. 14.6: The distribution of the dominant frequencies of 4-channel EGG signal
processing by the means of Butterworth filtering
Table 14.1: The NI values calculated for the exemplary of 4-channel (A1, A2, A3,
A4) EGG signal after the classical filtering
NI
Signal All Preprandial Meal Post prandial1 Post prandial2
channnel A1 0,8980 1,0000 0,5000 0,8158 1,0000
channnel A2 0,8163 0,7297 1,0000 0,8158 0,9412
channnel A3 0,8265 1,0000 0,3333 0,7632 0,7647
channnel A4 0,9694 1,0000 1,0000 0,9211 1,0000

Magnitude (dB) 0
normo = 0.908
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Magnitude (dB)
0
normo = 0.786
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Magnitude (dB)
0
normo = 0.786
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Magnitude (dB)
0
normo = 0.939
−50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Frequency (cpm)
Fig. 14.7: The distribution of the dominant frequencies of 4-channel EGG signal
processing by the means of the DCT filtering
Table 14.2: The NI values calculated for the exemplary of 4-channel (A1, A2, A3,
A4) EGG signal after the DCT filtering
NI
Signal All Preprandial Meal Post prandial1 Post prandial2
channnel A1 0,9082 1,0000 0,6667 0,8158 1,0000
channnel A2 0,7857 0,7568 0,5000 0,7895 0,9412
channnel A3 0,7857 1,0000 0,3333 0,7632 0,5294
channnel A4 0,9388 1,0000 1,0000 0,8421 1,0000
14.4 Discussion and Conclusions
In this paper the DCT filtering has been proposed as an alternative method for pre-
processing the raw EGG data. The obtained results confirm the good performance of
the proposed method. The pure EGG data has been extracted from their cutaneous
raw recording. The outcomes of the Butterworth and the DCT filtering obtained
for the human real EGG data are comparable (Table 33.1, 26.2). The DCT method
slightly influence the parameters of EGG data such as the dominant frequencies and
the normogastria index. The use of the DCT filtering allows more precise shaping
the amplitude-frequency characteristic of the EGG signal processing track, which
may affect the values of the parameters of the EGG examination. The benefit of the
such preprocessing the EGG signal can be found in the case, when the DCT filtering
is used for the noise cancellation in the highly interfered the EGG signals. The block
DCT filtering introduce an additional noise connected with the Gibbs phenomenon.
Our future work will be concentrated on extending the research by comparing the
proposed method with the other known methods (e.g. adaptive filtering) and exam-
ining the influence of the level of the noise to the efficiency of filtering the EGG
signal by means of the proposed method.
References
1. Alvarez, W.C.: The Electrogastrogram and What It Shows. Journal of the American Medical
Association, volume 78, 1922
2. Calder, S., O’Grady, G., Cheng, L. & Du, P., 2016.: A theoretical analysis of electrogastrog-
raphy (EGG) signatures associated with gastric dysrhythmias. IEEE Trans. Biomed. Eng., In
Press
3. Chen, J.D.Z, McCallum, R.W.: Electrogastrograpthy: measurement, analysis and prospective
applications. Medical & Biomedical Engineering & Computing, volume 29, pp.339-350, 1991
4. Hang Sik Shin, Chungkeun Lee, Myoungho Lee., Ideal Filtering Approach on DCT Domain
for Biomedical Signals: Index Blocked DCT Filtering Method (IB-DCTFM), Springer Science
+ Business Media, LLC, DOI 10.1007/s10916-009-9289-2, 2009.
5. Koch, K.L., Stern, R.M.: Handbook of Electrogastrography. Oxford University Press, 2004
6. Levanon, Daniela., Chen, J.Z.: Electrogastrography: Its Role in Managing Gastric Disorders
Journal of Pediatric Gastroenterology & Nutrition 27(4):431-443, October 1998
7. Medtronic A/S., Polygram NetTM Reference Manual. Skovlunde (Denmark), 2002.
8. Oppenheim, Alan V., Ronald W. Schafer, and John R. Buck., Discrete-Time Signal Processing.
2nd Ed. Upper Saddle River, NJ: Prentice Hall,(1999).
9. Parkman H.P., Hasler W.L., Barnett J.L., Eaker E.Y.: Electrogastrography: a Document Pre-
pared by The Gastric Section of The American Motility Society Clinical GI Motility Testing
Task Force. Neurogastroenterology and Motility, Vol. 15, 2003, pp. 89-102
10. Riezzo G., Russo F., Indrio F., Electrogastrography in Adults and Children: The Strength,
Pitfalls, and Clinical Significance of the Cutaneous Recording of the Gastric Electri-
cal Activity. BioMed Research International, vol. 2013, Article ID 282757, 14 pages,
doi:10.1155/2013/282757,(2013).
11. Tomczyk A., Jonderko J., Multichannel electrogastrography as a non-invasive tool for eval-
uation of the gastric myoelectrical activity a study on reproducibility of electrogastrographic
parameters before and after a meal stimulation. Ann. Acad. Med. Siles., 61, 5, 2007.
12. Yin, J., Chen J.D.Z.: Electrogastrography: Methodology, Validation and Applications, Journal
of Neurogastroenterology and Motility. 2013;19(1):5-17. doi:10.5056/jnm.2013.19.1.5
Chapter 15
Determining Heart Rate Beat-to-Beat from
Smartphone Seismocardiograms: Preliminary
Studies
Szymon Sieciński1 and Paweá Kostka1
Abstract. In the last decade the development of high quality, sensitive and inex-
pensive accelerometers has been seen, which in combination with low cost com-
putational power provided the reasons for reconsidering analysis of cardiovascular
vibrations in clinical practice.
Seismocardiography (SCG) is a non-invasive method of analyzing and recording
vibrations generated by heart activity and blood motion. Mobile devices offer the
possibility to monitor health parameters. Various applications have been proposed
for SCG, including HRV (heart rate variability) analysis. Our aim is to determine
location of AO points of SCG to achieve heart rate (HR) changes in time.
Proposed algorithm consists of calculating total acceleration value, signal pre-
processing, peak nding, computing time between consecutive AO peaks and con-
verting to heart rate. Algorithm performance was measured as true positive (TP),
false positive (FP), false negative (FN) rates, sensitivity (Se) and positive predictive
value (PPV) on 833 beats collected from 4 subjects.
We achieved average Se = 0.868 and PPV = 0.737 and in the best case Se =
0.995 and PPV = 0.974.
The obtained results are encouraging and indicate the possibility of measuring
heart rate beat-to-beat accurately in rest conditions.
Keywords: seismocardiography, heart rate variability, AO detection, smartphone
15.1 Introduction
In the last decade the development of high quality, sensitive and inexpensive ac-
celerometers has been seen, which in combination with low cost computational
Silesian University of Technology, Department of Biosensors and Biomedical Signals Processing,

Zabrze, Poland,
e-mail: pkostka@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_15
134 Paweá Kostka et al.
power provided the reasons for reconsidering analysis of cardiovascular vibrations

in clinical practice [11].
Seismocardiography (SCG) is a non-invasive method of recording and analyzing
vibrations generated by heart activity and blood motion, which are transmitted to the
chest wall [3]. Recordings are performed on subjects supine with the accelerometer
placed on the sternum, using the ultra-low frequency acceleration transducer with
linear response from 0.1 to 800 Hz and a sensitivity of 1.0 V/g (g = 9.81 m/s2 )
[9, 10]. In the literature, the majority of studies on SCG consider only analyzing the
dorso-ventral component. 3D SCG can be achieved by using tri-axial accelerometers
[1]. In this study three orthogonal components X, Y, and Z are used to calculate total
acceleration value.
SCG signal consists of the following points: mitral valve closure (MC), aortic
valve opening (AO), onset of rapid enjection (RE), aortic valve closure (AC), mitral
valve opening (MO), the peak of rapid lling (RF) and atrial systole (AS) [10].
Mitral valve opening has the highest amplitude [8]. Figure 15.1 presents a SCG
cycle with annotated ducial points.
Seismocardiogram is a non-stationary signal and there are morphology variations
between subjects. In particular, the AO peak may appear to have two neighboring
peaks. This is the issue which complicates developing annotation algorithms [11].
Another issue concerning SCG annotation is the susceptibility to subject’s motion,
respiration and noises [8]. We consider primarly AO peak as the element of SCG
signal with the highest amplitude.
Fig. 15.1: SCG vs. ECG by Ghufran Shaq et al. Image retrieved from [8]. License:
CC-BY 4.0.
Part (a) shows raw SCG signal (above) and ECG signal (below). Part (b) presents
annotated SCG and ECG ensemble averages (dark lines) and superimposed SCG
and ECG beats (light shades).
Various applications have been proposed for SCG, including heart rate variability
analysis, detecting heart arrhythmias, myocardial ischemia [1, 2, 11].
15 HRV Analysis using Smartphone SCG 135
Our aim is to detect AO points to determine heart rate variability (HRV) using
seismocardiograms acquired solely by smartphone.
15.2 Methods
15.2.1 Experiment protocol
The experiment was conducted on 4 healthy subjects (with no heart disease diag-
nosed) aging 54 ± 16.8077 years. Each subject was lying down in supine position
at rest. Each SCG measurement lasted 193.09-203.42 seconds and was performed
using LG H440n smartphone, with sampling frequency fs = 100 Hz and sensitiv-
ity of 0.009 m/s2 . The smartphone was placed loosely on sternum according to
gure 15.2. We used Accelerometer Meter mobile application available on Google
Play to register precordial accelerations.
Fig. 15.2: Smartphone placement and coordinate system.
15.2.2 HRV Estimation Algorithm
Proposed algorithm consists of three stages: calculating the total acceleration value,
preprocessing, peak nding and calculating heart rate (HR).
Peak locations indicate the AO points. Heart rate is calculated as in equation 15.1
60
HR(t) = , (15.1)
tAO−AO
where tAO−AO is the AO-AO period in seconds.
15.2.3 Signal Processing
The SCG signal registered by a smartphone consists of three orthogonal components

X, Y, Z. Signals were exported to text les and analyzed off-line using MATLAB
software (The Mathworks, Inc., Natick, MA, USA).
In the rst step total acceleration magnitude is calculated according to the equa-
tion 15.2.
v(t) = x(t)2 + y(t)2 + z(t)2 , (15.2)

where x(t), y(t) and z(t) represent the acceleration components X, Y and Z
against time respectively. The use of equation 15.2 was proposed by Migeotte et
al [5].
Total acceleration signal v(t) was band-pass ltered using FFT with cut-off fre-
quencies of 0.5 and 40 Hz. In order to remove artifacts, upper signal power envelope
is calculated from the v(t) using the root mean square (RMS) operation. A sliding
window with a length of 0.5 s was applied and an artifact threshold was dened as
twice the median value of upper power envelope. Signal values above the threshold
were not considered in the further analysis.
In order to simplify signal structure, an analytical envelope with a FIR lter
length equal to 50 samples was calculated from ltered v(t) signal. Then, we used
a FIR running average lter of length N = 8. Equation 15.3 describes the running
average lter.
1 N
y[n] = ∑ bi x[n − i],
N i=0
(15.3)
where bi is the i-th element the lter coefcients vector, y[n] the output signal
and x[n] is the input signal. Then, the signal is band-pass ltered using FFT lter
with cut-off frequencies of 0.5 and 3 Hz.
15.3 Results
We considered 4 recordings of seismocardiograms acquired from 4 subjects. All the

analyzed recording were annotaded manually to analyze performance of proposed
algorithm.
Table 15.1 contains subjects description. Table 15.2 shows number of beats in
each analyzed signal, true positives (TP), false positives (FP), false negatives (FN),
sensitivity (Se) and positive predicitve value (PPV). True positive is dened as the
AO point detected correctly. False negative is considered if proposed algorithm
omits the AO point and false positive is determined for misclassied AO points.
Sensitivity (Se) is dened as Se = T P+FN

TP
and positive predictive value (PPV) is
dened as PPV = T P+FP . The number of beats is the sum of TP and FN.
TP
Figure 15.3 presents an example of AO point detection results on subject No. 3

and gure 15.4 presents heart rate variability (HRV) graph generated with proposed
algorithm. Presented HRV graph has four peaks which indicate the inuence of
false positive beats near the “true” beats. Local maxima of processed signal are not
always indicators of beats due to the high variability of SCG morphology [6, 7].
SCG with detected AOs

0.3
0.2
2
]
0.1
acceleration [m/s
-0.1
-0.2
32 34 36 38 40 42
time [s]
Fig. 15.3: Example of detection results on subject No. 1. Asterisks indicate detected
AO points.
Table 15.1: Subjects description.
Subject number Age [years] Gender Recording length [s]

1 62 Male 193.09
2 63 Female 203.48
3 25 Male 203.42
4 66 Female 203.33
HR(t)
200
180
160
140
HR [bpm]
120
100
80
60
40
0 50 100 150 200
t [s]
Fig. 15.4: Heart rate againts time graph generated using the proposed algorithm for
subject No. 3.
Table 15.2: Number of beats in analyzed signals and algorithm performance param-
eters.
Subject number TP FP FN Beats Se PPV

1 164 36 27 191 0.859 0.820
2 186 145 54 240 0.775 0.562
3 184 5 1 185 0.995 0.974
4 183 126 34 217 0.843 0.592
total 717 312 116 833 0.861 0.697
average 179 78 29 208 0.868 0.737
PPV rates are lower in all subjects than sensitivity due to higher FP rate than FN
rate.
Sensitivity and positive predictive value of algorithm on male subject record-
ings is higher than on female subject recordings due to lower rate of false posi-
tives caused by the smartphone size and the steady placement of recording device.
Movent activities which appeared during the recording affect the quality of analyzed
signals due to peaks classied as AO points.
For all annotated 833 beats, the sensitivity and positive predictive value is highly
affected by FP and FN rates of female subjects. The results are slightly higher than
of female subjects.
Proposed algorithm performs best on SCG signal No. 3 acquired from a 25-
year old male subject due to the constant position of smartphone and staying still
during the experiment. The lowest results were achieved on subject No. 2 due to its
movements and smartphone placement during the experiment.
15.4 Discussion
We performed the measurements on very limited number of subjects in resting con-

ditions supine and assessed the performance of algorithm using the manually anno-
tated AO points. Determining heart rate using AO-AO period depends on reliable
detection of AO points. Considerably lower PPV than Se indicates the need of re-
ducing false positives, while maintaining high sensitivity.
Due to lower algorithm performance on female subjects, more suitable smart-
phone position should be considered, as observed Landreani et al. [4]. Performance
measures values for male subject supine seem encouraging and indicate the possi-
bility of determining heart rate beat-to-beat accurately.
Next studies should include more subjects, comparing SCG signals registered by
smartphone and high precision accelerometer, processing time measurements and
applying robust noise and artefacts removal algorithms.
References
1. Inan OT, Migeotte PF, Park KS, Etemadi M, Tavakolian K, Casanella R, Zanetti J, Tank J,
Funtova I, Prisk GK, Di Rienzo M: Ballistocardiography and seismocardiography: a review
of recent advances. IEEE J Biomed Health Inform. 2015 Jul;19(4):1414-27.
2. Jafari Tadi M et al: A real-time approach for heart rate monitoring using a Hilbert transform
in seismocardiograms. Physiol. Meas. 37 1885 (2016)
3. Korzeniowska-Kubacka I: Sejsmokardiograa — nowa nieinwazyjna metoda oceny czynności
lewej komory w chorobie niedokrwiennej serca. Folia Cardiol. 2003, tom 10, nr 3 265-268.
4. Landreani F et al.: Beat-to-beat heart rate detection by smartphone’s accelerometers: Valida-
tion with ECG, 38th Annual International Conference of the IEEE Engineering in Medicine
and Biology Society (EMBC), Orlando, FL, 2016, pp. 525-528.
5. Migeotte P-F, Tank J, Pattyn N, Baevsky R, Neyt X and Prisk G: Three dimensional ballis-
tocardiography: methodology and results from microgravity and dry immersion. Conf. Proc.
IEEE Eng. Med. Biol. Soc. 4271-4 (2011).
6. Paalasmaa J, Toivonen H and Partinen M 2014 Adaptive heartbeat modelling for beat-to-beat
heart rate measurement in ballistocardiograms IEEE J. Biomed. Health Inform. 19 1945–52
7. Rienzo M D, Vaini E, Castiglioni P, Merati G, Meriggi P, Parati G, Faini A and Rizzo F 2013
Wearable seismocardiography: towards a beat-by-beat assessment of cardiac mechanics in
ambulant subjects Autonomic Neurosci. 178 50–9
8. Shaq, G. et al. Automatic Identication of Systolic Time Intervals in Seismocardiogram. Sci.

Rep. 6, 37524; doi: 10.1038/srep37524 (2016).
9. Zanetti JM and Salerno DM: Seismocardiography: A Technique for Recording Precordial Ac-
celeration. Proceedings of Fourth Annual IEEE Symposium of Computer-Based Medical Sys-
tems, Baltimore, MD, USA, 1991, pp. 4-9.
10. Zanetti and Salerno DM: Seismocardiography: waveform identication and noise analysis.
Computers in Cardiology, Venice, Italy, 1991, pp. 49-52.
11. Zanetti JM, Tavakolian K: Seismocardiography: Past, Present and Future. Proceedings of 35th
Annual International Conference of the IEEE EMBS, Osaka, Japan, 3 - 7 July, 2013.
Chapter 16
Field modelling of several electromagnetic drive
variants of the micropump blood transfusion
device
Sebastian Bartel, Zbigniew Pilch, and Tomasz Trawiński
Abstract. In this article the trends in constructing and modelling of devices sup-
porting the blood flow in the human heart will be presented. The latest Eurotrans-
plant statistics report on the incidence of heart failure and the latest trends in the
treatment of the failing heart will be presented in chapter one. Trends in developing
ventricular assist devices and the number of implanted device types supporting the
dysfunctional heart will be discussed in chapter two. Finally there will be shown a
modelling method of the electromagnetic system of pulsatile artificial hearts with
the research results.
Keywords: Heart failure; micropump; artificial heart; VAD
16.1 Introduction
The European Society of Cardiology (ESC) reports that around 26 million adults
worldwide are suffering from heart failure and cardiovascular disease. The inci-
dence of heart failure in industrialized countries is about 2% among all adults and
from 6% to 10% incidence in the case of adults over 65 years old. Cardiovascu-
lar disease is very common in the group of people between 60-79 years old (70%).
Only in the USA the mortality rate from cardiac diseases alone was about 0.8 million
deaths in 2010. Depending on a region and degree of industrialization, from 17%
to 45% of the patients die within the first year and the others die within the next 5
years. The best method offered by the medical community today for the treatment
of heart failure is a heart transplant. Unfortunately, the low number of hearts for
transplantation makes it impossible to treat all the patients that are waiting for the

e-mail: {sebastian.bartel,zbigniew.pilch,tomasz.trawinski}@polsl.pl,
WWW home page: http://www.polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_16
142 Tomasz Trawiński et al.
organ [3-6]. Eurotransplant statistics report that the highest rate of patients wait-
ing for heart transplant occurred between 2012-2013 and amounted to about 1200
cases. In 2015 the ratio was still high, for the active waiting list it was 1170 patients
and the available transplant organs equalled 605. In 2015 the number of heart trans-
plantations in Poland was close to 100 operations, whereas in the same year heart
transplantations in Germany exceeded the number of 250 operations [4]. The article
presents the current issue of the incidence of acute heart failure in the United States
and Europe.
In the situation of high mortality caused by heart failure and the insufcient num-
ber of hearts for transplantation, it is necessary to work on ventricular assist devices
(VAD) and total articial hearts (TAH).
16.2 Trends in developing ventricular assist devices and total

articial hearts
The examples of various ventricular assist devices presented in this article are char-
acterized by their dimensions, blood ow and energy requirements. All the LVAD
devices from the rst project to the newest project achievements can be divided into
ve generations of cardiac devices. This division is presented on the example of
HeartAssist devices generation that is designed and manufactured by ReliantHeart
Incorporated.
Table 16.1: LVAD devices generation type.
Generation type Name Weight Origin

1 Novocor 1 kg Pulsatile pump
2 HeartMate 0.28 kg Continous axial ow pump
3 Levacor 0.4 kg Rotary pump with non-contact bearing
4 HVAD 0.145 kg Centrifugal rotary pump with magnetic bearings
5 HeartAssist > 0.1 kg Rotary pump with low energy consumption and low
mass
As one can see in the table above, the characteristic feature of LVAD devices in
the newest generations is to lower their mass and energy consumption while main-
taining the same blood ow. The latest fth generation of LVAD devices uses tech-
nologies with magnetic levitation bearings and optimally shaped inlets and outlets
of the rotor so as to minimize the formation of thrombosis. Thrombi are especially
dangerous when they detach from the device and get into the bloodstream block-
ing blood vessels and leading to venous congestion, haemorrhage, and in the worst
cases, death of a patient. To avoid such incidences LVAD devices evolved from large
and heavy rst-generation pulsatile pumps to light and small centrifugal ones spe-
cialized in supporting the left ventricle. Centrifugal pumps of this type, for example
Heart Assist 5 device, are characterized by rotary speed between 7000 and 12000
Title Suppressed Due to Excessive Length 143
RPM and blood ow from 2-100 L/min. Small dimensions of devices supporting
the work of the heart have the advantage that an operation on a patient is minimally
invasive and has a low number of unsuccessful operations. The disadvantage of this
solution is a laminar blood ow resulting in a low hemodynamic pulse in the pa-
tient‘s blood circulatory system. When the pulse is weak it must be helped by the
supporting device. The next group of devices presented in this article is total arti-
cial heart devices. In contrast to LVAD devices TAH ones are developed in order
to completely replace the human heart. We can present the following examples of
TAH devices:
1. AbioCor is a totally implantable TAH device that replaces the human heart.
The total mass of the pump is about 1 kg and it is armed with a pneumatic drive
system that creates the pressure for the blood ow.
2. Syncardia is a pneumatic driven biventricular and orthotopic implantable total
articial heart with the weight of 160 g. The blood ow rates in the Syncardia
TAH device are between 7 - 9 L/min.
3. The Carmat pneumatic TAH device consists of two ventricular blood cham-
bers. Each of the two chambers is separated by a exible bio-membrane, one
for blood and one for the actioning uid. Carmat generates a blood ow up to
10 L/min and is equipped with an internal power supply.
4. POLTAH is a completely inside the body implantable pneumatic TAH device
designed by Polish bioengineers. The main circuit of blood ow is separated
from the pneumatic circuit by a exible polyurethane membrane and the whole
construction is based on crystalline titanium and its compounds.
All of the above devices are based on a pneumatic drive using compressed air or
a special uid as the pulsatile work medium of the articial heart. In these construc-
tions the blood chamber is divided into two parts separated by a exible membrane.
This is done to separate two circuits from each other: the main blood circuit and
the pneumatic circuit. The volume of the chamber and the blood outlet volume are
changed by the volume of the pneumatic chamber. The volume characteristic of the
blood chamber is a very important hemodynamic parameter and it is described as
the cardiac output (CO). This parameter is determined by the average blood ow
volume which is transfused through the human blood system in the course of one
minute [12]. Therefore, it is very important to monitor the volume of the blood
chamber and the output blood volume. An example of the blood chamber volume
determination method is the optical measurement system, which uses a matrix of
several light sources and light detectors placed in the pneumatic chamber. This type
of method is described in detail in the article “Optoelectronic system for the deter-
mination of blood volume in pneumatic heart assist devices“ [11].
Fig. 16.1: The conceptual pulsatile micropump and its ve-section electromagnetic
drive.
16.3 Geometrical modelling of electromagnetic pulsatile articial

heart
The electromagnetic drive is a linear motor based on the technology of voice coil
motors (VCM) and is placed in the upper part of the pulsatile micropump. The gure
below (Fig. 16.1) shows an example of computer visualization of the geometry of
the micropump that represents a single articial blood chamber and its drive. The
proposed geometrical model of the articial heart consists of a hydraulic system
and electromagnetic drive. For a better visualization the upper permanent magnet
and the magnetic yoke were removed in this gure. The inlet/outlet of the single
blood chamber is denoted by number (1) and its diameter is equal to the diameter of
the aorta (28 mm). The bearing in this construction is denoted by (3) and connects
the movable portion of the rotor with the housing. The blood chamber consists of
a exible biocompatible membrane and is denoted by (2). The maximum span of
the articial chamber and rotation angle of the rotor is 45 degrees. The maximum
momentary volume (75 ml) of the chamber occurs at the angle of the rotor equal
to 45 degrees and the minimal volume occurs at 0 degree. The stator windings of
the drive are denoted by (4) and the rotor with the permanent neodymium magnet
denoted by (5). The stator windings are divided into ve sections from S1 to S5
each sections consists of three coils controlled independently. In the micropump’s
current solution the stator windings are placed on an arc that has a circle of 120 mm
radius. The permanent magnet system moves on an arc above and under the stator
windings, where the centre of rotation is in the middle of the bearing.
The following gure (Fig. 16.2) shows a simplied idea of a micropump drive
with a neodymium permanent magnet rotor and stator windings, where the colours
Fig. 16.2: Deceptive picture showing section S1 and S2 of conceptual drive of blood
pump
indicate the next pair of active sides of the individual stator coils. The electro-
magnetic drive is parametrized so the parameter wm is the width of the permanent
magnet and thus the width of the single section, while the parameter hm is the height
of the magnet. Parameters related to the stator winding are wc - the width of the sin-
gle active side of the coil, hc - the height of the active side, and parameter g denotes
the width of the air gap between the rotor magnets and it is strongly correlated with
the permanent magnet working point.
Then, for the adopted concept of magnetic circuit there were assumed dimensions
for which a key limitation was the availability of permanent magnets in sale. Other
dimensions were parametrised for the dimensions of the magnets available in stores.
The dimensions of the magnets, coils and air gap are shown in Fig. 16.2, and the
dimensions for the 6 accepted cases are summarized in Table 2.
Table 16.2: Characteristic dimensions adopted as variables in electromagnetic com-

putations
Case Magnet Magnet Air gap Coil Coil

width wm height hm g width height
wc hc
1 15 2.5 4 2.3 3.6
2 21 2 4 3.3 3.6
3 25 5 6 3.9 4.6
4 61 5 11 9.8 4.6
5 37 5 10 5.8 5
6 20 2 4 3 2
In addition, it is assumed that the width of the winding is 1/3 of the width of
the magnet (wc = wm /3). ). The simulation model was developed in the FEMM
environment and for all six simulated cases the model depth was the same and was
equal to 30 mm. In the 2D model there were generated 35997 nite elements with
the total number of nodes 18263. The calculations were performed by moving the
coils in the x-axis from the initial position x = 0 to the end position x = wm , with
step dx = 0.5mm. Then the gathered magnetic induction values are shown as the
characteristic of BAV = f (x). The symmetrical reection of the value 0 corresponds
with the values of the magnetic induction, assuming that the displacement is in the
negative x direction.
Fig. 16.3: Magnetic ux density distribution of one selected case of magnetic drive
circuit
Fig. 16.4: Averaged magnetic induction in the middle of the air gap for the six cases
of magnetic circuit congurations listed in Table 2.
Figure 3a shows the eld model of the selected number of 6 drive cases with
stator windings and permanent rotor magnets formed on an arc shape as can be seen
in Fig. 16.1. Figure 3b shows the same drive case but the eld model is simplied
so that the stator windings and permanent rotor magnets are rectangle-formed. This
simplication is intended to speed up the eld modelling of different geometries
Fig. 16.5: Generated force values as a function of rotor position for the six consid-
ered micropump drive congurations.
of the micropump drive. The difference of the generated forces between the two
modelling types of the electromagnetic drive is negligibly small. For the arc type
model the simulated force is equal to 1.4406 N, while the generated force for the
rectangle type model is equal to 1.4435 N. For these two models the simulations
were made for a zero rotor position, where the generated force is the maximum.
Figure 3c shows the rotor of the drive in its maximum position, where the generated
force is minimal.
The simulated magnetic induction distribution of the considered drive cases al-
lows to determine the forces acting on the actuator rotor. In the FEMM model the
stator winding was tted with AWG32 (American Wire Gauge) wire material of 0.2
mm diameter without insulation and 0.24 mm diameter with insulation. For calcu-
lations, the value of the current owing through stator windings was 200 mA. The
family of characteristics BAV = f (x) for the six cases listed in Table 2 are shown in
Fig. 16.4. The results for the simulated force for all drive geometric cases are shown
in Fig. 16.5.
16.4 Conclusions
Performed numerical calculations for several variants of the magnetic circuits of the
prototype of pulsatile micro pump drive system allow to formulate the following
conclusion:
• It is possible to build a magnetic circuit of the pulsatile micropump drive system
in the assumed overall dimensions.
• Based on the assumed dimensions of the magnetic circuit (Fig. 16.1),the process
of optimizing the dimensions of the magnetic circuit components can be carried
out to reduce the dimensions and mass while maintaining the approximate force
generated by the drive.
• The analysed electromagnetic actuators (6 variants described in Table 2) give
different magnetic ux densities which differ greatly from one another (Fig. 16.3).
• For the assumed dimensions of the articial blood chamber and for the nec-
essary blood volume (85ml), displacement of the actuator by the distance of
95mm is required. This means that 2 or 5 sections are required for the tilting.
• In order to realize the displacement of the actuator, it is necessary to supply 6
coils at the same time. The current that ows through these coils does not need
to have the same value at a given time, as is apparent from the control algorithm.
• Considering the electromagnetic propulsion of the micropump on a perimeter
model allows for much faster optimization of the dimensions of this drive. For
the proposed six microprocessor drive cases, the generated force shown in the
gure represents a wide range of maximum forces of 3 to 17 N. In a further
consideration it is necessary to dene the criterion of the maximum force that
is needed to generate the blood pressure of 120 mmHg and the stroke volume
of approximately 70 ml.
References
1. Bartel, S., Trawiński, T.: Design issues of electromagnetic micropump. 15th International
Workshop on Research and Education in Mechatronics (REM), 9-11 September 2014, El
Gouna, Egypt. Piscataway: Institute of Electrical and Electronics Engineers, pp. 1–7, (2014).
2. Bartel, S., Trawiński, T.: The inuence of the micropump‘s winding shape and magnetic cir-
cuit conguration on the generated electromagnetic torque characteristic. Part I: FEM analysis,
Przeglad Elektrotechniczny, R92, pp. 240–245, (2016).
3. Bartunek, J., Vanderheyden, M.: Transplantation approach to heart failure. Springer Scie-
nce+Business Media New York, (2013).
4. http://statistics.eurotransplant.org
5. http://my.americanheart.org
6. http://www.vascularmodel.org/miccai2013/
7. Wenyu Shi, Meng-Sang Chew: Mathematical and Physical Models of a Total Articial Heart.
2009 IEEE International Conference on Control and Automation Christchurch, New Zealand,
December 9-11, pp. 637–642, (2009).
8. Shoaib, M.S.R., Haque, M.A., Asaduzzaman, M.: Mathematical Modeling of the Heart. 6th In-
ternational Conference on Electrical and Computer Engineering ICECE 2010, 18-20 December
2010, Dhaka, Bangladesh, pp. 626–629, (2010).
9. Yalincinkaya, F., Kizikaplan, E., Erbas, A.: Mathematical Modelling of Human Heart as a Hy-
droelectromechanical System, 8th International Conference on Electrical and Electronics Engi-
neering (ELECO), pp. 362–366, (2013).
10. Goldoozian, L. S., Zahedi, E.: Mathematical modeling of heart rate and blood pressure varia-
tions due to changes in breathing pattern. Proceedings of 20th Iranian Conference on Biomed-
ical Engineering (ICBME 2013), University of Tehran, Tehran, Iran, December 18-20, pp. 54–
58, (2013).
11. G. Konieczny, T. Pustelny, M. Satkiewicz, M. Gawlikowski, „Optoelectronic system for the
determination of blood volume in pneumatic heart assist devices” Konieczny et al. BioMed Eng
OnLine (2015).
12. M. Gawlikowski, T. Pustelny, R. Kustosz, M. DarÅĆak: Non invasive blood volume mea-
surment in pneumatic ventricular assist device POLVAD. Fundation for Cardiac Sugery De-
velopment, Zabrze Poland, Department of Optoelectronics, Silesian University of Technology,
Gliwice Poland, Molecular and Quantum Acoustics vol. 27, (2006).
Chapter 17
Minimizing interference from cardiac stimulator
pulse in the ECG recordings during the
diagnostics of myocardial ischemia by
non-invasive transcutaneous cardiac stimulation
Jerzy Gałecka1 , Fryderyk Prochaczek1 , and Adam Gacek1 ,

and Hanna Winiarska-Prochaczek2
Abstract. Paper describes new method minimizing ECG records interference dur-
ing non-invasive transcutaneous cardiac stimulation allowing wider use of this
method for diagnosis of myocardial ischemia. By moving the electrodes of ECG
precordial leads to lower intercostal spaces it was achieved the smaller penetration
of pulse from front surface stimulating electrode. Eliminating the possibility of in-
put circuit overload by interference has improved the effects of applied solutions
of stimulating pulses suppression during the test. Waveforms of shifted leads are
shown in the records after conversion of successive samples to a value correspond-
ing to standard position on the basis of coefficients being determined for each patient
at the test beginning. Achieved have been better records in all patients and in some
patients eliminated was the problem of collision occurrence between standard loca-
tion of precordial leads and optimal location of the front stimulating electrode thus
impeding or preventing execution of the test.
Keywords: noninvasive transcutaneous cardiostimulation, ECG signal, interference

suppression
17.1 Introduction
Execution of classical electrocardiographic stress test in patients with symptoms

suggestive of coronary artery disease is not always possible. Such a situation we
encounter in patients with intermittent claudication, damage to the musculoskeletal
system, high blood pressure, or limitation of exercise capacity (NYHA III-IV). In
this case, the dobutamine test is taken out or the patient is referred to the hemody-
namic laboratory in order to perform the coronary angiography.
Institute of Medical Technology and Equipment ITAM, Zabrze, Poland,

e-mail: jerzy.galecka@itam.zabrze.pl,
WWW home page: http://www.itam.zabrze.pl · Cardiologic Outpatient Clinic CAR-
DIOTEST, Tychy, Poland

https://doi.org/10.1007/978-3-319-70063-2_17
152 Jerzy Gaáecka et al.
Similar situation is in the patient hospitalized due to chest pain, in which pain
subsides after admission to the hospital, and normal ECG together with negative
troponin level tests rule out myocardial necrosis. Obtainment of assuredness as to
the cause of the pain requires observation of the patient in hospital conditions and
repeating the checkups during a possible recurrence of pain, or to transfer a patient
without reasonable grounds to hemodynamic lab. In both of these situations, a ben-
eficial alternative may be to perform stimulatory coronary test using transcutaneous
non-invasive cardiac stimulation.
This test can also be useful for the screening of at-risk population to identify
potential risks of coronary heart disease. The advantage of this test is a short cardiac
workload duration of 30 seconds, and the possibility of obtaining the maximum
pulse rate limit (calculated basing on the age of the subject). This is not assured
by other methods, including coronary test performed with transesophageal atrial
stimulation [1], [2]. Due to the fact that the current transesophageal method from
the standpoint of applicable law is classified as invasive technique, it is used to
perform stimulation coronary test through diagnostics run at a different angle, for
example: detection of undocumented previously arrhythmia.
Transcutaneous cardiac stimulation has been in use for over 50 years in the car-
diac therapy [3]. Its intensive development, especially for diagnostic applications,
occurred in the 70s and 80s of the twentieth century, with significant participation
of Polish researchers [4], [5], [6]. Ventricular stimulation is possible, because ven-
tricles are closer to the front surface of the chest than the atria. So far, numerous
attempts have been undertaken to optimize the methods of heart stimulation, in or-
der to obtain a better tolerance of examination in the patient [7], [8] as well as
gaining clear stimulation records from its course, desired by the physician [9].
Effective and well-tolerated, longer lasting stimulation of the heart ventricles via
the transcutaneous way in a standard 2-electrode system is obtained in about 50 %
of patients [10]. This follows from the low threshold of stimulation noticing (on av-
erage 10 mA) and a relatively high individual excitation threshold, of 30 to 100 mA
with a pulse duration of 30 ms. At higher thresholds the stimulation can cause ex-
citation of skeletal muscles of the chest. Tolerance of short-duration transcutaneous
stimulation is also an individual matter and is in the range of 50 to 120 mA. Proposed
by the authors in 1994, a new 3-electrode system of transcutaneous stimulation has
reduced the discomfort of stimulation [8] and has enabled the introduction of tran-
scutaneous stimulation coronary test for wider use. In this embodiment, the lower
threshold values of ventricular excitation and lesser skeletal muscle stimulation is
obtained by placing the active electrode on the front of the chest, in the electrocar-
diographic point C3 and two passive electrodes in the cross-gluteal vicinity on both
sides of the spine.
Coronary transcutaneous stimulatory test (CTST) is only the 30 seconds lasting
strain on the heart with stimulation accelerating its rate of contractions to 130 or
150 bpm. During the test, the patient remains in the supine position. His skeletal
muscles of the lower limbs are at rest, and therefore there is no formation of skeletal
muscle oxygen debt, as it happens in a classic stress test on a treadmill or cycle
ergometer. The results of the authors’ studies in healthy volunteers have excluded
17 Minimizing interference from cardiac stimulator pulse 153
significant influence of 30-second transcutaneous stimulation of heart ventricles on

morphology of the ST-T interval (ie. electric memory of the heart), which is the
basis of electrocardiographic evaluation of ischemia [11].
17.2 Sources of interference
Non-invasive coronary stimulation test, just like the classic exercise test, for the as-
sessment of results requires recording a full 12-lead ECG. In the case of stimulation
test performed, it is hampered due to occurrence in the record very large artifacts
from the cardiac stimulator pulse. Typical stimulating pulse of non-invasive exter-
nal cardiac stimulator, effectively exciting the heart ventricles, has energy approx.
10 000 times higher than is the case for intracardiac stimulation. The elimination of
this interference was a big challenge for the designers.
The interference from the stimulus pulses appear at the input of patient’s signal
acquisition channel together with the useful signal, the source of which is the heart.
A useful signal, received from the chest and limbs, has substantially the voltage am-
plitude in the range of 0.5–3.0 mV. In contrast, recorded together with the useful
signal in surface leads interferences from the cardiac stimulator pulses far exceed
the scope of the A/D processing in the signal acquisition channels of modern elec-
trocardiographs or ECG monitors. Sloping exponentially response of the impedance
of electrodes system to cardiac stimulator pulse reaches the values close to useful
signal only after 100–200 milliseconds, especially in the precordial leads V2, V3,
V4, V5 (Fig. 41.1).
Fig. 17.1: Example response to stimulating pulse at the input of acquisition channel
of ECG from precordial leads in the standard arrangement.
This is due to unfavourable for the received signal quality the standard leads
location near the front active stimulating electrode optimally placed (Fig. 41.2).
Fig. 17.2: Typical arrangement of ECG precordial leads and front stimulating elec-
trode.
Sometimes, however, it occurs that the optimal from the point of view of the
stimulation threshold position of the front stimulating electrode interferes with the
standard location of the ECG electrodes of mentioned leads in the fourth intercostal
space. This greatly impedes or prevents performing the examination well tolerated
by a patient, or causes complete disruption of records. So it became necessary to
find a solution ensuring good recordings quality at optimum position of the front
stimulating electrode in each patient.
On the poles of surface ECG electrodes are also contact voltages, whose ampli-
tude and rate of change depend on the material used and the stability of conditions
occurring at the place of contact with the skin. This necessitates the use of high-
pass filtering and optionally stabilizing isoelectric line. In therapeutic applications,
the cutoff frequency of high-pass filter is mostly 0.5 Hz. However, for the proper
interpretation of records in the course of a diagnostic exercise stress test required is
the cut-off frequency of component at the level of 0.05 Hz.
Transcutaneous stimulation may result in additional exciting the skeletal muscles
and therefore the movement of chest surface, which is a source of isoelectric line
instability in waveforms from precordial leads. Whereas, muscle contraction in re-
sponse to started stimulation may be a source of interference visible primarily in the
limb leads.
17.3 Methods of interference suppression
Until the nineties of the twentieth century in electrophysiological laboratories used

were high-class analog electrocardiographs. They had a wide range of overdrive,
and additional function of accelerating the isoelectric line return, used in the course
of non-invasive transcutaneous stimulation by pulses of large amplitude. This was
making possible to obtain a good quality ECG immediately after the stimulation
allowing diagnostic interpretation of ischemic changes in the ECG.
In order to ensure the possibility of evaluation of the effectiveness and the course
of conducted stimulation in all conditions all the non-invasive cardiac stimulators
developed ITAM had included a built-in ECG acquisition channel equipped with
the necessary systems of filters and interference suppression of stimulating pulses
[8]. Similar developments were also used in therapeutic cardiac stimulators of other
companies in the world.
Analog cardiac stimulator of first generation NP-4D (1991), as well as digitally
controlled cardiac stimulator of second generation NAP-601 with built-in monitor
(1995), were fitted with a single ECG signal acquisition channel. It contained the
high-pass and low-pass analog filters and suppression system consisting of ana-
log keys cooperating with a high-pass filter. This system for the programmed sup-
pression time maintained input potentials at the output before the stimulator pulse
appeared, preventing overload of cooperating electrocardiograph or ECG monitor.
These solutions were sufficient for therapeutic applications.
Coronary transcutaneous stimulation test (CTST) was performed initially us-
ing non-invasive cardiac stimulator and specialized electrocardiograph Cardis-210
(ITAM) (Fig. 33.3), equipped with a hybrid (analog-digital) system for detection
and suppression of interference from pulses [12], [13].
Fig. 17.3: Workstation to perform coronary transcutaneous stimulation test consist-

ing of the non-invasive cardiac stimulator NAP-601 (ITAM) as well as specialized
electrocardiograph Cardis-210 (ITAM).
For diagnostic evaluation of test results used was the piece of writing made before
stimulation start and the fragment after turning off stimulation and stabilization of
isoelectric line (Fig. 33.4).
Introduced by ITAM on the market in 2010 diagnostic and therapeutic device Di-
agnoStim MDD-501 integrates non-invasive cardiac stimulator as well as the ECG
acquisition channel and waveform memory of 12 surface leads. It is thus possible
execution of CTST without additional devices at the patient’s bedside, as well as
in cooperation with a standard electrocardiograph through the analog output at a
milivolt level. This is to print records both during and after the test.
Fig. 17.4: Fragment of the record carried out before stimulation (A) and a record
fragment from the final stimulation phase and after it is turned off (B). Visible sig-
nificant interference from pulses S1 (despite the applied suppression) and significant
ST segment depressions in leads V5, V6, indicating the occurrence of ischemia.
DiagnoStim can also transmit the ECG waveforms in digital form to the ECG
computer system CARDIV (ITAM), both during and after the test (Fig. 33.5). In
parallel signal is sent in the form filtered by low- and high-pass filters in the MDD-
501, as well as in the form of raw samples to a more advanced filtration in the
computer application CARDIV.
The acquisition channel of ECG signal from 12 surface leads of DiagnoStim
device has a high resistance to overload resulting from the use of 22-bit SigmaDelta
converter. This allows us to eliminate artifacts originated from the cardiac stimulator
pulses, from the waveforms recorded after processing them into digital form. To
cooperating electrocardiograph or monitor in analog way the waveforms are sent,
being free from pulse interference, so the conventional devices may be used, not
having a high resistance to overdrive and the interference. Cooperation with the
electrocardiographic computer system CARDIV is carried out via optical link as a
standard USB. This ensures high electrical safety of a patient and a high level of
electromagnetic compatibility required for medical systems.
Fig. 17.5: Workstation to perform coronary transcutaneous stimulation test consist-

ing of multifunction device MDD-501 (ITAM) as well as computer electrocardio-
graph CARDIV (ITAM).
17.4 Minimizing interference in the precordial leads
Occurring in some patients collision of standard arrangement of precordial ECG

electrodes and optimal positioning of the front active stimulating electrode (–) pre-
vents or significantly hinders to perform the test. Too close location of the electrodes
causes a penetration of stimulator pulse of amplitude up to a maximum of 80 V di-
rectly to the ECG electrodes, which completely disorganize records of all leads.
Moving the stimulating electrode causes most frequently the significant increase
in the threshold of stimulation, worsening the patient’s tolerance of test. Therefore
analyzed and checked was the concept of shifting the ribbon of precordial electrodes
by one or even two intercostal spaces below and conversion so recorded signals
using the coefficients being calculated.
Records of precordial leads of the 4th, 5th and 6th intercostal space were per-
formed in a group of 45 patients with normal resting ECG. It was found that
recorded R-wave amplitude was decreased by 30 % for 5th and 50 % for the 6th
intercostal space on average. Considerable individual differences in measurement
results between patients did not allow, however, apply the solution with the adop-
tion of fixed coefficients for each lead based on the average results of measurements
in the study group. Therefore, adopted solution is to calculate the coefficients in-
dividually for the particular patient basing on results of amplitude measurement of
QRS complexes in the individual leads for the standard position and shifted before
the start of stimulation.
We also assessed the impact of position of precordial leads ribbon on the size
of emerging ischemic changes in ECG. Made in 6 patients with a positive CTST
result simultaneous record of two precordial ECG leads indicating the deepest ST-
segment depression in the 4th, 5th and 6th intercostal space showed the changes of
ST segment depression proportional to changes in R-wave amplitude (Fig. 33.6).
Fig. 17.6: On example recording of coronary stimulation test can be seen the change
of ST depression in leads V4 and V5 deployed in IV (a), V (b) and VI (c) intercostal
space in proportion to the change in amplitude of R wave.
It seems therefore, that this method can reproduce ECG waveforms, which were
received during CTST in standard leads, based on the waveforms from leads in
the 5th or 6th intercostal space, without losing information about the reduced ST
segment indicating the stimulation induced myocardial ischemia (Fig. 17.7).
In order to adopt this solution for wide use, accepted by the medical community,
it is necessary to carry out studies in larger groups of patients with a positive result of
CTST. It is also necessary to verify the diagnostic results of CTST using diagnostic
and therapeutic invasive proceeding, that is coronary angiography and stenting of
narrowed vessels.
Fig. 17.7: Fragment of the record carried out before stimulation start (A) and a
record fragment from the final stimulation phase and after it is turned off (B) for
electrodes V3–V6 shifted to 6 intercostal space. In part (B) of the record, visible
small interference from pulses S1 and essential ST segment depressions in leads
V3–V6, indicating the occurrence of myocardial ischemia.
17.5 Summary
The solution presented helps minimize interference from the stimulating pulses in
precordial leads during coronary transcutaneous stimulation test (CTST) in the pres-
ence of collision between a standard location for precordial electrodes in 4th inter-
costal space and optimal position of the active front stimulating electrode from the
point of view of stimulation threshold. This allows us to expand the group of patients
that can be diagnosed by this method, as well as to obtain better quality recordings
in all examined patients using this method as a standard for recording precordial
lead during CTST. The described solution is used in the system CardiScreen de-
signed to conduct non-invasive screening to assess the ris of cardiovascular disease
occurrence, particularly at higher risk group — men over 40 years of age.
References
1. Schröder, K., Wler, H., Dingerkus, H., MÃijnzberg, H., et al.: Comparison of the Diagnos-
tic Potential of Four Echocardiographic Stress Tests Shortly After Acute Myocardial Infarc-
tion: Submaxima Exercise, Transesophageal Atrial Pacing, Dipyridamole, and Dobutamine-
Atropine. Am J Cardiol. 77, pp. 909–914 (1996)
2. Zoll, P.M., Linenthal, A.J., Norman, L.R., Paul, M.H., Gipson, W.: External Electric Stimu-
lation of the Heart in Cardiac Arrest. A. M. A. Archives of Internal Medicine, pp. 639–653
(1955)
3. Prochaczek, F., Gałecka, J.: New clinical applications of non-invasive transcutaneous cardiac
pacing. (Chapter 14 in: Non-invasive transcutaneous cardiac pacing, edited by Birkui P.I.,
Trigano J.A., Zoll P.M. Futura Publising Company Mount Kisko, New York, pp. 161–178
(1992)
4. Prochaczek, F., Gałecka, J., Jarczok, K.: Programmed non-invasive transcutaneous cardiac
pacing: examination of the A-V retrograde conduction. RBM, 14, pp. 311–313 (1992)
5. Prochaczek, F., Birkui, P.J., Surma, I., Gacek, A., Jarczok, K.: Sinoatrial conduction tested by
transcutanous cardiac stimulation. RBM, 16, pp. 124–127 (1994)
6. Zoll, P.M., Zoll, R.H., Belgard, A.H.: External non-invasive electric stimulation of the heart.
Critical Care Medicine Vol. 9, No.5, pp. 393–394
7. Prochaczek, F., Birkui, P.J., Gałecka, J., Jarczok, K.: Is the new electrode configuration a break
point in transcutaneous cardiac pacing tolerance?. RBM 16, pp. 98–101 (1994)
8. Prochaczek, F., Gałecka, J.: The effect of suppression of the distortion artifact during transcu-
taneous pacing on the shape of the QRS complex. PACE 13, pp. 2022–2025 (1990)
9. Prochaczek, F., Kowalska, I., Orzeł, A., Owczarek, A., Gałecka, J.: Evaluation of electrical
cardiac memory induction during non-invasive coronary stimulation test. Proceedings of the
XI International Conference Medical Informatics and Technology MIT 2006, edited by Pi˛etka
E., Ł˛eski J., Franiel S., pp. 145–150 (09.2006)
10. Luck, J.C., Markel, M.L.: Clinical applications of external pacing: a renaissance? PACE, Vol.
14, pp. 1299–1316 (08.1991)
11. Prochaczek, F., Gałecka, J., Skowronek, A.: Evaluation of an analog system for spike arte-
fact elimination in ECG records during transcutaneous stimulation of heart ventricles. Prace
Naukowe Instytutu Górnictwa Politechniki Wrocławskiej Nr 86. Seria: Konferencje Nr 24.
IMEKO TC-10 Conference on: Integration in Technical Diagnostics. Wrocław, pp. 212–217
(09.1999)
12. Prochaczek, F., Gałecka, J., Gibiński, P., Skowronek, A.: The meaning of ECG-signal filtration
for correct interpretation of recordings performed during diagnostic and therapeutic cardios-
timulation. Journal on Medical Informatics and Technologies, University of Silesia, Vol. 2,
Part II, pp. MT–41–47 (2001)
13. Prochaczek, F., Gałecka, J.: Non-invasive electrostimulation technologies for the diagnosis
and therapy of coronary artery disease and heart arrhythmias. Journal of Medical Informatics
and Technologies, Vol. 9, pp. 25–34 (2005)
Chapter 18
Novel tumor protein markers collection by the
use of highly porous organic material for the
upper and lower respiratory system –
preliminary results
Andrzej S. Swinarew1 , Barbara Mika2 , Jarosław Paluch3 Jadwiga Gabor1 , Marta

Ł˛eżniak1 , Hubert Okła1 , Tomasz Flak1 , Beata Swinarew4 , and Klaudia Kubik1
Abstract.
In the era of XXI century, when modern imaging techniques allows to increase
detection of upper and lower respiratory tract cancer, still almost 75% of patients are
diagnosed at an advantage stage. A very important problem in diagnostic practice is
a general trend to use routine methods and the lack of new techniques exploration.
Nowadays laboratory diagnosticians known defects of used methods which result
limited effectiveness. Despite this fact too little attention is paid to the confrontation
of biomedical and physicochemical views what follows to develop new methods
within the framework of interdisciplinary research.
Late detection of cancer changes is often associated with inability to take effec-
tive treatment. Most modern measurement techniques based on liquid chromatogra-
phy combined with mass spectrometry gives opportunity to detect millions particles
of substances. These techniques are being used by scientists from all over the world
to create database of proteins, which are characteristic for almost all pathological
changes.
Keywords: protein markers, cancer, cancer diagnostics
Institute of Materials Science, Faculty of Computer Science and Material Science, University of
Silesia in Katowice,
e-mail: andrzej.swinarew@us.edu.pl · Department of Biosensors and Biomedical Sig-
nals Processing,
Silesian University of Technology, Zabrze, Poland
· Department and Clinic of Laryngology, School of Medicine in Katowice,
Medical University of Silesia in Katowice,
Francuska 20/24, 40-027 Katowice, Poland
· Institute for Engineering of Polymer Materials and Dyes,
Paint and Plastics Department,
44-100 Gliwice, Poland

https://doi.org/10.1007/978-3-319-70063-2_18
162 A.S. Swinarew et al.
18.1 Introduction
Cancer is one of the most common causes of death. In the lead of fatal types are
lung, bronchus, breast, colon, rectum, prostate and pancreas cancers [1].
Expansion of most tumors proceeds almost asymptomatic, however early detec-
tion increases chances of positive therapeutic result. Even those patients who were
diagnosed with lung cancer in a first of advancement stage, have at most around
60% probability of 5-year survival, which means that in a large proportion of these
patients already occurring clinically detectable metastases [2].
The biggest problem encountered by doctors is fact, that tumors of upper and
lower respiratory system are particularly difficult to detect in early stage. Mainly
because symptoms are very similar to many minor illnesses. That is why in addition
to research on the treatment of advanced cancer, scientist have to pay special atten-
tion to preventive diagnostics, which as methods of early detection combined with
ability to treat cancer can multiple the survival of patients.
In detecting presence of cancer helps tumor markers. For individual tumor there
are few specific markers. In most cases they are non-specific, this is one marker
may be associated with the occurrence of various cancers (tumor-associated mark-
ers). Ideal biomarker should unambiguously characterize important from the clinical
point of view attributes of the tissue such as tumor growth, stage, sensitivity or resis-
tance to treatment. It should be also characterized by high reliability of diagnostic,
i.e. the corresponding sensitivity and specificity determination and do not require
invasive procedures for collecting biological materials and high cost performance
indications [3].
There are five main types of markers possible for screening chemical analysis:
• Enzymes;
• Hormones;
• Embryonic and fetal antigens;
• Tumor-associated mucin glycoproteins antigens;
• Genetic markers (detection of oncogenic factors and gene mutations) [1].
The big hope of laboratory diagnosticians raise chemical factors which are mark-
ers of neoplastic proliferation. A soluble hydrophilic and lipophilic compounds
(plasma components) as well as gaseous components, which can be measured in
exhaled air, are constantly being sought in case of the respiratory and digestive sys-
tems tumors.
The composition of exhaled air is very rich in substances which may be produce
in metabolic diseases. Among the potential factors which may be gaseous tumor
markers are mentioned: isoprene, aniline, o-toluidine and branched alkanes [4, 5].
An important group of tumor markers are also fragments of DNA and RNA circulat-
ing in peripheral blood [6]. In the literature were described DNA fragments which
were identified by real-time PCR as markers for esophageal cancer [7].
A demonstration of the protein presence is direct evidence of cellular activity, in
contrast to the studies related to genome expression, which potentially may have
18 Novel tumor protein markers collection. . . 163
no functional significance. Proteomic analysis which is focused on methods for de-

tection of tumor markers, is a very important element in the diagnosis of cancers
[8].
Modern molecular techniques permit efficient and reliable tests of any mutation
or expression of genes and proteins in tumor tissue [9].
The basic diagnostic materials still remains tissue from primary tumor gained
with invasive methods (surgical resection, puncture through the wall of bronchi or
chest). Clearly, this imposes restrictions on the wider and more effective use of
molecular diagnostics in clinical practice. Meanwhile, it is already know that lung
cancer-specific markers are present not only in the genetic material of cells obtained
from the tumor itself, but also in the sputum, Broncho alveolar lavage fluids and
blood serum or plasma of patients [10].
Tumor markers (Fig. 18.1) are the cellular products released into body fluids
or present on the cell surface. Most of them don’t have to accompany malignant
changes, which means that they are not specific. However they detection may fa-
cilitate the diagnosis and monitoring of disease, since they are produced by various
types of normal and cancer cells.
Because of the differentiating factors markers can be classified into prognostic
and predictors types.
First of these helps to choose method of adjuvant therapy, eligible patients at
high risk of relapse, optionally direct to adjuvant therapy and help in the decision to
discontinue treatment of patients with good prognosis.
Predictive tests are mainly to help in the selection of therapy based on molecular
phenotype of tumor sensitivity, and qualify for the topical treatment of patients.
Prognostic test results are subject to specific comparisons with standards to which
they must meet certain guidelines.
It is important that the results were confirmed clinically, and their assessment
should be reproducible, widely available with guaranteed quality control and easy
to interpret.
Testing levels of tumor markers is generally simple and to perform it we need
only patient blood. The most commonly assayed markers:
• CEA;
• M2-PK;
• CYFRA 21-1;
• NSE;
• SCC;
• Beta-HCG (germ cell tumors, trophoblastic disease) [11].
18.2 Methods
18.2.1 Synthesis of sorbent material
It is known that potassium glycidoxide in the presence of 18-crown-6 oligomer-

izes spontaneously forming cyclic products with several alkoxide active centers
(Fig.18.1) [12]. These oligomers are used as macroinitiators for the polymeriza-
Fig. 18.1: Synthesis of multifunctional oligoethers
tion of oxirane monomers. Star-shaped polyethers with a crown-like core and with
mainly three or six hydroxyl end groups are obtained this way (Fig.18.2) [13].
Fig. 18.2: Structure of obtained star shaped cyclic cores
In this work we show a new route to star-shaped poly (propylene oxide) with four
arms and four hydroxyl end groups.
18.2.2 MALDI-TOF analysis
The synthesis was performed at 25o C in a 50 cm3 reactor equipped with a magnetic
stirrer and teflon valves enabling the substrates delivery and sampling under argon
atmosphere. Into the reactor containing a weighed amount of potassium hydride (5
mmol) a tetrahydrofuran solution (8 cm3 ) of glycidol (5 mmol) and propylene ox-
ide (20 mmol) was added and the mixture was stirred 20 min. Then, 18-crown-6
(5 mmol) was very slowly dropped into the reactor. After monomer consumption
Dowex 50WX2 ion exchanger was introduced into the reaction mixture to trans-
form the alkoxide groups into the hydroxyl end groups. The molecular mass of
poly(propylene oxide) macromolecules was estimated by means of MALDI-TOF-
TOF technique on a Shimadzu AXIMA instrument. The spectra were obtained in a
positive ion-mode. Norharmane was used as the matrix.
Three sets of peaks are seen in the MALDI-TOF spectrum of poly(propylene oxide)
obtained (Fig.18.3). Peaks of the high intensity belong to set marked A and peaks
of the low intensity belong to sets B and C.
The analysis of the spectrum (Table 34.1) shows that set A represents potassium
adduct ions with the macromolecules containing four propane-1, 2-dioxide units
(G), which are formed from potassium glycidoxide, and four OH end groups. Sets
B and C represent potassium adduct ions of macromolecules with three or five G
units and three or five OH groups, respectively.
Fig. 18.3: MALDI-TOF spectrum of poly(propylene oxide)
Potassium glycidoxide formed in the reaction of potassium hydride with glycidol

during the first step of the process is an inimer. It means it can act both as alkoxide
Table 18.1: Experimental and calculated masses of poly(propylene oxide) macro-

molecules
Potassium Number of m/z Mn

Series
adduct ions ∗ units experimental calculated
x=4, y=7 740,00 741,64
A Gx (PO)y Hx K + x=4, y=7 799,00 799,72
x=4, y=9 858,34 857,80
x=3, y=8 725,19 725,76
B Gx (PO)y Hx K + x=3, y=9 783,94 783,78
x=3, y=10 842,53 841,86
x=5, y=6 757,44 757,58
C Gx (PO)y Hx K + x=5, y=7 815,99 815,66
x=5, y=8 874,63 873,74
initiator and as oxirane monomer. However, it initiates its self-oligomerization or the

polymerization of propylene oxide only after activation by 18-crown-6, and it as the
monomer is more reactive than propylene oxide. Thus, the polymerization under
study can start only if 18-crown-6 is added to the reaction mixture and activated
potassium glycidoxide attaches mainly own molecules.
Taking into account the results of earlier works [14, 15, 16] it is assumed that the
cyclization occurs when the growing oligomer contains six ether oxygen atoms be-
cause this oligomer surrounds closely the potassium counter-ion similarly as crown
ether does. The alkoxide end group can react in this case with the oxirane ring of
potassium glycidoxide incorporated into the chain as the initiator. The cyclic hex-
amer is formed chiefly from four potassium glycidoxide molecules and two propy-
lene oxide ones. The arms of this cyclic oligomer start to growth when successive
amounts of 18-crown-6 are introduced into the reaction mixture. Finally, star-shaped
polyether with the cyclic core of an irregular structure and with four arms is ob-
tained. However, the macromolecules with three and five arms are also observed.
In the age of medical science rapid progress it is not difficult to note new paths
that can facilitate diagnosis and speed up the detection of potential killer. In recent
years we are noticing growing popularity of exhaled air composition analysis. This
situation is due to development of analytical chemistry, including gas chromatogra-
phy (GC) and mass spectrometry(MS).
Non-invasive breath test using polymer sampling bags (Fig. 18.4) is one of the
cheapest and easiest way to detect cancer markers. The principle of operation is
simple, patient blows air into the bag, then bag with breath sample goes to analyse.
However, this method suffers from a fairly serious problem of transport filled
bags. Inflated like a balloon becomes very cumbersome and sensitive to even the
slightest damage. Therefore, many centers around the world are working on a
method to replace them.
One of such method is to use sorption porous material (Fig. 18.5) in the form of
polyurethane foam, which thanks to extensive internal structure are able to close on
much smaller space protein markers.
Fig. 18.4: Gas sampling bag Fig. 18.5: Sample of poliurethane sorp-
tion porus material
By creating material that occupy less space and being resistant to mechanical
damage, we are facing the problem of deposition searched markers. In case of sam-
pling bags patients had to blow them once. When we consider foam it is necessary
to repeat these steps several times to deposit the appropriate amount of markers.
Therefore, in parallel with studies of sorbents, scientists are trying to get medical
holder allows to filter out air pollutants inhaled by patients, and direct exhaled breath
on presented foam.
The use of experimental FDM 3D printer we managed to get first reusable holders
that can be autoclaved at temperatures exceeding 200o C.
18.4 Conclusions
Main role of marker prevalence research is early disease diagnosis and search for
new ways of treatment, which promotes the survival of patients with cancer of upper
and lower respiratory tract.
Research on the use of exhaled air, as a factor in order to determinate state of
patient health, because of low cost and non–invasive method of performing is one
of the most promising diagnostics methods.
Ongoing work on the use of polyurethane foam to replace gas sampling bags
allow to save space with similar production cost.
Usage of foams extends possible storage time of samples due to lack of protein
consumption by bacteria, which is caused by size of the pores in the internal struc-
ture of foam.
Reliable picture of cancer prediction should include list of more than one test
result bears witness that indicates presence of increased levels tumor markers in
exhaled air. Every individual tumor have few specific markers, so one marker may
be associated with the occurrence of various cancers. Most of them does not have
to be accompanied by malignant changes, but their detection can help in the rapid
diagnosis of early stage cancer or in decision to discontinue treatment.
So far scientists were not succeeded in fully explanation of the mechanism re-
sponsible for release tumor markers into circulatory system. These markers may
occur in healthy and diseased patients. Concentration of tumor markers serum may
indicate whether patient is sick or if there’s increased risk of developing one of
cancers.
Individual markers are not specific, so their presence can not determinate occur-
rence of cancer. Concentrations of markers in blood vary not only due to disease
risk, but also dependent on age and sex of test subject.
References
1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Smigal C, Thun MJ. Cancer Statistics. CA Cancer
J Clin 56: 106-30.
2. Patz EF. Jr, Campa MJ, Gottlin EB, Kusmartseva I, Guan XR, Herndon JE. Panel of Serum
biomarkers for the diagnosis of lung cancer. J Clin Oncol. 2007; 25: 5578-83.
3. Brambilla C, Fievet F, Jeanmart M, de Fraipont F, Lantuejoul S, Frappat V, Ferretti G, Brichon
PY, Moro-Sibilot D. Early detection of lung cancer: role of biomarkers. Eur Respir J. 2003;
21: 36-44.
4. Poli D, Carbognani P, Corradi M, Goldoni M, Acampa O, Balbi B, Bianch L, Rusca M, Mutti
A. Exhaled voatile organic compounds In patient with non-small cell lung cancer: cross sec-
tional and nested short-term follow-up study. Respir Res. 2005; 6: 71-81.
5. Horvath I, Lazar Z, Gyulai N, Kollai M, Losonczy G. Exhaled biomarkers in lung cancer. Eur
Respir J. 2009; 34: 261-75.
6. Fleischhacker M, Schmidt B. Circulating nucleic acids (CNAs) and cancer-A survey. Biochim
Biophys Acta. 2007; 1775: 181-232.
7. Tomita H, Ichikawa D, Ikoma D, Sai S, Tani N, Ikoma H, Fujiwara H, Kikuchi S, Okamoto
K, Ochiai T, Otsuji E. Quantification of circulating plasma DNA fragments as tumor markers
in patients with esophageal cancer. Anticancer Res. 2007; 27: 2737-42.
8. Conrad DH, Goyette J, Thomas PS. Proteomics as a method for early detection of cancer:
a review of proteomics, exhaled breath condensate, and lung cancer screening. J Gen Intern
Med. 2007; 23: 78-84.
9. Skroński M, Szpechciński A, Chorostowska- Wynimko J. Current methods to detect EGFR
gene mutations as predictive factor for targeted therapies in non-small cell lung cancer – is
there a “golden standard” in diagnostics? Pneumonol Alergol Pol. 2014; 82: 311-22.
10. Chrostowska-Wynimko J, Skroński M, Szpechciński A. Molekularne markery prognostyczne
i predykcyjne w diagnostyce niedrobnokomórkowego raka płuca. Onkologia Info. 2011; 3:
160-7.
11. Schneider J, Velcovsky HG, Morr H, Katz N, Neu K, Eigenbrodt E. Comparison of the tumor
markers tumor M2-PK, CEA, CYFRA 21-1, NSE and SCC in the diagnosis of lung cancer.
Anticancer Res. 2000; 20: 5053-5058.
12. Stolarzewicz A, Morejko-Buż B, Grobelny Z, Pisarski W. Spontaneous self-oligomerization
of potassium glycidoxide – A simple way to new cyclic polyfunctional macroinitiator. React
Funct.Polym. 2005; 65: 259-66.
13. Morejko B, Stolarzewicz A, Grobelny Z, Piekarnik B, Niedziela T, Trzebicka B. New kind of
star-shaped polyethers prepared with cyclic oligo(potassium glycidoxide) as a macroinitiator.
React Funct Polym. 2007; 67: 669-74.
14. Stolarzewicz A, Morejko-Buż B, Grobelny Z, Pisarski W, Lanzendörfer M, Müller A. Struc-

ture of poly(propylene oxide) obtained with potassium glycidoxide in the presence of crown
ether. Rapid Commun Mass Spectrom. 2004. 18: 716-20.
15. Kazanskii KS. Donor-acceptor and solvation interactions in anionic polymerization of some
heterocycles. Pure Appl Chem. 1981; 53: 1645-61.
16. Arkhipovich GN, Dubrowskii SA, Kazanskii KS, Ptitsina NV, Shupik AN. Study of solvation
of alkali cations with poly(ethylene oxide). Eur Polym J. 1982. 18: 569-76.
Chapter 19
Statistical Analysis of the Impact of Molecular
Descriptors on Antimicrobial Activity of
Thiourea Derivatives Incorporating
3-amino-1,2,4-triazole Scaffold
Anna Filipowska1,2 , Wojciech Filipowski1 , Ewaryst Tkacz2 , Monika Wujec3
Abstract. The paper contains results of statistical and chemometric analysis for 15
thiourea derivatives containing the 3-amino-1,2,4-triazole moiety and characterized
by antimicrobial activity against Staphylococcus aureus (NCTC 4163, ATCC 25923,
ATCC 6538, ATCC 29213), Staphylococcus epidermidis (ATCC 12228) bacteria, as
well as by low cytotoxicity (or lack thereof) against infected MT-4 cells. Multiple
regression and cluster analysis were employed to perform the study. The research
enabled obtaining linear relationships connected with three molecular descriptors
SA, η , log P. The conducted chemometric analyses indicate that the increase in
activity against the studied strains is closely related to the type and position of sub-
stituent in a phenyl ring.
Keywords: Statistical analysis, cluster analysis, antimicrobial activity, thiourea

derivatives, 3-amino-1,2,4-triazole
19.1 Introduction
Chemicals containing a thiourea moiety are among important chemicals modified

in order to obtain derivatives with the most beneficial biological, including antibac-
terial, properties. In the last decade, a lot of attention was paid to derivatives with
heterocyclic 1,2,4-triazole moiety as potential pharmacological agents. Scientific re-
search shows that a 1,2,4-triazole ring bound to coumarin [1], naphthyridine [2] or
indole [3] is a structurally important part of compounds with anticancer activity. Ad-
Faculty of Automatic Control, Electronics and Computer Science, Silesian University of

Technology, Poland, e-mail: afilipowska@polsl.pl,
WWW home page: · Department of Biosensors and Processing of Biomedical Signals, Silesian
· Department of Organic Chemistry, Medical University of Lublin, Poland

https://doi.org/10.1007/978-3-319-70063-2_19
172 Anna Filipowska et al.
ditionally, triazoles and their derivatives are worth attention due to their wide spec-
trum of antifungal [4] and antibacterial [5, 6, 7] activity. In order to minimize costs
of traditional in vitro screenings employed in design of new biologically active com-
pounds, they are enhanced by computer in silico research based i.e. on correlation
between molecule’s chemical structure and a defined process with its participation.
This procedure is known as the QSAR (Quantitative Structure-Activity Relation-
ship) method. This method is based on the assumption that the differences between
particular compounds in terms of antimicrobial activity result from differences in
their chemical structure.
This paper concerns determination of physicochemical parameters of 15 disub-
stituted thiourea derivatives (Table 19.1) with the 3-amino-1,2,4-triazole moiety, de-
scribed in a paper by Joanna Stefańska et al. [8, 9]. These derivatives are character-
ized by moderate antimicrobial activity against Staphylococcus and low cytotoxicity
(or lack thereof) against infected MT-4 cells. The determined molecular descriptors
and antimicrobial activity against Staphylococcus epidermidis and Staphylococcus
aureus bacteria were utilized as basis for chemometric analyses and attempts at cre-
ating QSAR models describing relationships between antimicrobial activity against
analyzed strains and compound structure based on molecular descriptors.
19.2 Experimental Works and Methods
Molecular descriptors describing physicochemical properties of thiourea derivatives

incorporating 3-amino-1,2,4-triazole moiety (Table 19.1) were determined.
Physicochemical parameters for all 15 thiourea derivatives were determined by
means of the HyperChem software, version 7.5 [10]. Conformational analysis was
performed by means of the OPLS force field method with default conformational
space search settings. Next, the most stable structures (having the lowest energy)
were optimized by means of the PM3 semi-empirical method. The applied gradient
value is 0.01 kcal/mol·Å [11].
19 molecular descriptors have been determined for each analyzed compound: M
— mass, log P — partition coefficient octanol/water [12], SAg — surface area —
grid, SA — surface area — approx., V — volume, Rf — molar refractivity, α —
polarizability, μ — dipole moment, HOMO — the highest occupied molecular or-
bital energy, LUMO — the lowest unoccupied molecular orbital energy, HLG —
difference between HOMO and LUMO energy levels, η — hardness described with
the following formula: η = (LUMO - HOMO)/2 [13], χ — Mulliken electronega-
tivity expressed by the following equation: χ =-(LUMO + HOMO)/2, ET — total
energy, EB — binding energy, EIA — isolated atomic energy, EE — electronic en-
ergy, IC−C — core-core interaction, HF — heat of formation. Next, a chemometric
analysis was performed, by employing cluster analysis (CA), in order to classify the
analyzed chemical compounds based on their chemical composition and physico-
chemical properties. Data standardization was the first stage of cluster analysis. Its
purpose is to increase the influence of data with minor deviation and decrease the
19 Statistical Analysis of the Impact of Molecular Descriptors. . . 173
Table 19.1: Structures of thiourea derivatives used in the study.

Compound number R1
1 C6 H5
2 4-Cl-C6 H4
3 4-F-C6 H4
4 2-Br-C6 H4
5 4-Br-C6 H4
6 4-I-C6 H4
7 2-F-C6 H4
8 3-F-C6 H4
9 2-Cl-C6 H4
10 3-Cl-C6 H4
11 3-Cl-4-CH3 -C6 H3
12 3-(CF3 )-C6 H4
13 3-Cl-4-CH3 -C6 H4
14 3-Cl-4-F-C6 H3
15 2-COOC2 H5 -C6 H4
influence of data with major deviation. Furthermore, the standardization procedure

eliminates the influence of different units of measurement and renders the data di-
mensionless [14]. Analysis of similarities was performed by means of the Ward’s
method with Euclidean distance metric. Statistical analysis was also performed in
order to determine linear correlations between antimicrobial activity and molecu-
lar descriptors. The statistical analysis was conducted using the STATISTICA 12
software [15]. Linear relationships between antimicrobial activity and structural de-
scriptors were determined by means of multiple linear regression. Values of mini-
mal inhibitory concentration (MIC [μ g/ml] presented in Table 19.2), expressed as
log(1/MIC), for 5 bacterial strains (Staphylococcus aureus NCTC 4163, Staphylo-
coccus aureus ATCC 25923, Staphylococcus aureus ATCC 6538, Staphylococcus
aureus ATCC 29213, Staphylococcus epidermidis ATCC 12228) were used as de-
pendent variables, while structural parameters determined in the course of research
were used as independent variables.
Coefficients of linear equation were calculated by means of the classic least
squares method. The statistical calculations were conducted at confidence level of
95% (p < 0.05). The analyses were conducted by means of multiple backward re-
gression which entailed successively rejecting the least statistically significant (hav-
Table 19.2: Antimicrobial activity of 15 thiourea derivatives against 5 bacterial

strains (Staphylococcus aureus NCTC 4163, Staphylococcus aureus ATCC 25923,
Staphylococcus aureus ATCC 6538, Staphylococcus aureus ATCC 29213, Staphy-
lococcus epidermidis ATCC 12228) expressed as minimal inhibitory concentration
(MIC [μ g/ml])
Number of S.aureus S.aureus S.aureus S.aureus S.epidermidis

compound NCTC4163 ATCC25923 ATCC6538 ATCC29213 ATCC12228
1 100 100 100 100 50
2 25 25 25 25 25
3 50 50 50 25 25
4 50 50 50 50 25
5 50 50 50 25 25
6 25 25 25 25 25
7 200 200 200 200 200
8 50 50 50 50 50
9 32 16 32 32 16
10 16 8 8 8 8
11 16 8 8 8 8
12 8 4 8 8 4
13 32 32 64 64 16
14 16 8 16 16 8
15 32 16 32 16 16
ing the highest p value) structural parameters (molecular descriptors). Statistical

verification of the model was performed by means of T-test and Fisher-Snedecor F-
test. T-test is used to verify significance of each parameter of a model (independent,
explanatory variables), and F-test is used for checking statistical significance of the
entire model.
After statistical verification, the model was checked for compliance with the co-
incidence axiom and verified in relation to multiple regression assumptions (i.e.
model linearity, homoscedasticity condition, normal distribution of model residue).
The model residue normality assumption was verified by means of the Shapiro-Wilk
test. Every independent variable used in the QSAR model is related to at least 5
compounds [16]. Activity-structure linear relationships were validated by means of
the Leave-One-Out Cross Validation (LOO CV). Successively one compound was
eliminated from the data set (N) used as a basis for a model and utilized to validate
the resulting new model containing N − 1 elements. The Q2 validation coefficient
was calculated from the formula 19.1 [17, 18].
2
∑ni=1 yexp,i − y pred,i
Q = 1−
2
2
(19.1)
∑ni=1 (yexp,i − yave,i )
where: yexp − log(1/MIC) value with index i obtained from measurements (ex-
perimental), y pred − log(1/MIC) value with index i, calculated based on a model
(predictor variable), yave - average value calculated for (N − i) compounds.
19.3 Results
Fig. 19.1 presents grouping results for 15 analyzed thiourea derivatives and stan-
dardized molecular descriptors describing them. Fig. 19.2 presents analogous group-
ing results, but with compound numbers substituted with values of their average an-
timicrobial activity (expressed in [μ g/ml]) against all analyzed Staphylococcus au-
reus (NCTC 4163, ATCC 25923, ATCC 6538, ATCC 29213) and Staphylococcus
epidermidis ATCC 12228 strains. The graphical representations of grouping results
presented below are included to facilitate interpretation of analysis of similarities
graphs.

%
#

$
#

"
!

Fig. 19.1: Grouping results for 15 analyzed thiourea derivatives and standardized
molecular descriptors describing them. X axis — names of molecular descriptors;
Y axis — numbers of analyzed compounds.
Fig. 19.3 presents a dendrogram showing division of analyzed chemical com-

pounds based on their antimicrobial activity for all 5 analyzed strains, expressed as
compound minimal inhibitory concentration MIC (μ g/ml). As visible in the den-
drogram, by applying the Sneath’s criterion, compounds may be divided into three
groups. The first group contains compound no. 7 characterized by antimicrobial ac-
tivity MIC = 200 [μ g/ml]. The compound incorporates a 2-fluorophenyl substituent
in the R1 position. Group III contains compounds no. 13, 8, 4, 5, 3 and 1 character-
ized by average antimicrobial activity from MIC = 40 [μ g/ml] to MIC = 90 [μ g/ml].
The group contains compounds incorporating in position 1 of thiourea moiety a
phenyl ring substituted in different positions with a Br or F atom. This group also

!

Fig. 19.2: Grouping results for 15 analyzed thiourea derivatives and standardized
molecular descriptors describing them. X axis — names of molecular descriptors;
Y axis — numbers of analyzed strains.
contains compound no. 13 with 4-(trifluoromethyl)phenyl in the R1 position. Group

II is characterized by the highest average antimicrobial activity from MIC = 6,4
[μ g/ml] to MIC = 25 [μ g/ml]. This group contains all compounds incorporating in
position 1 of thiourea moiety a phenyl ring substituted in different positions with a
chlorine atom and all compounds with phenyl rings substituted in two positions.
Fig. 19.4 presents a dendrogram with compounds grouped on the basis of SA,
Rf, log P, η , LUMO, HOMO descriptors describing steric, lipophilic and electronic
parameters. The analyzed compounds were divided into 5 groups designated with
letters A, B, C, D and E. Group A contains compound no. 15 incorporating in po-
sition 1 of thiourea moiety a phenyl ring with the 2-COOC2 H5 moiety substituted
in the ortho position. Group B contains compound no. 11 incorporating a phenyl
ring substituted with chlorine and fluorine atoms, as well as compounds no. 13 and
14 incorporating in the R1 position a phenyl ring with trifluoromethyl substituted in
the meta or para position. Group C contains compounds no. 11 and 6 with a phenyl
ring substituted with an atom of iodine or chlorine and methyl. Group D consists
of compounds no. 4, 5, 9, 10 and 2. Compounds incorporate in the R1 position a
phenyl ring substituted in different positions with an atom of bromine or chlorine.
The last group E contains compounds no. 7, 8, 3 and 1 with a phenyl ring substi-
tuted in different positions with a fluorine atom or not substituted at all. The group
is characterized by the lowest values of antimicrobial activity and, as presented in
fig. 19.1, low values of such descriptors as Rf, log P, SA. The performed analy-

"
"

!
!

" !
Fig. 19.3: Dendrogram of grouping of 15 thiourea derivatives based on antimicrobial

activity against all analyzed strains.

Fig. 19.4: Dendrogram of grouping of 15 thiourea derivatives based on antimicrobial

activity against all analyzed strains.
sis indicates that the substitutent in the a phenyl ring strongly influences values of
physicochemical parameters of studied compounds.
Relationships between antimicrobial activity and molecular descriptors for 5
Gram-positive bacterial strains Staphylococcus aureus (NCTC 4163, ATCC 25923,
ATCC 6538, ATCC 29213), Staphylococcus epidermidis (ATCC 12228) were de-
termined for all 15 derivatives. Values of minimal inhibitory concentration for an-
alyzed S. epidermidis and S. aureus bacterial strains are presented in table 19.2.
Modeling of quantitative structure-activity relationships (QSAR) was performed by
means of the commonly used multiple linear regression (MLR) [19] method. Mini-
mal inhibitory concentrations (MIC, μ g/mL) obtained for analyzed bacterial strains,
expressed as log(1/MIC) were used as dependent variables and molecular descrip-
tors were used as independent variables. Validation of the obtained QSAR models
was performed using the Leave-One-Out Cross Validation (LOO CV) method. The
equations were determined in such a way that each explanatory variable was related
to at least 5 compounds. Equations describing relationships between antimicrobial
activity of the 15 analyzed compounds and SA, η and log P molecular descriptors
were determined for the Staphylococcus aureus (NCTC 4163, ATCC 25923, ATCC
6538, ATCC 29213), Staphylococcus epidermidis (ATCC 12228) strains. It was not
possible to find statistically significant equations meeting all conditions for a larger
number of molecular descriptors. In the majority of cases the coincidence condition
was not met which is an indicator of equation redundancy and may imply that the
explanatory variables are linearly dependent. The obtained equations are presented
in Table 19.3.
As presented in Table 19.3, activity against the analyzed strains shows a positive
coefficient of correlation with SA, i.e. activity increases with the increase of surface
area of analyzed compounds. Similarly, antimicrobial activity also increases with
the increase of the η descriptor value expressed as (LUMO − HOMO)/2. HOMO
and LUMO descriptors describe electronic properties of compounds. According to
the frontier molecular orbital theory, compound reactivity may be determined on
the basis of interactions between HOMO and LUMO orbitals [20]. In the case of
Staphylococcus aureus (NCTC 4163, ATCC 25923, ATCC 6538), Staphylococcus
epidermidis (ATCC 12228) strains a statistically significant correlation between ac-
tivity and log P (parameter describing compound lipophilicity) is also noticeable.
Increase in log P is correlated with the increase in antimicrobial activity of the 15
analyzed compounds. This parameter describes the ability of chemical compounds
to permeate cell (biological) membranes.
Figures 19.5, 19.6 and 19.7 present predictive abilities of obtained equations
characterized by the highest correlation coefficients R as diagrams of predicted val-
ues in relation to observed (experimental) values for S. aureus and Staphylococcus
epidermidis strains.
Fig. 19.8 presents a dendrogram with compounds grouped on the basis of SA, μ ,
log P descriptors, for which linear relationships were found. The compounds were
divided into 4 groups designated with letters A, B, C and D. Group A contains com-
pound no. 15 incorporating in position 1 of thiourea moiety a phenyl ring with the
2-COOC2 H5 moiety substituted in the ortho position. This compound is character-
Table 19.3: Equations describing the relationship between antimicrobial activity

against Staphylococcus aureus (NCTC 4163, ATCC 25923, ATCC 6538), Staphylo-
coccus epidermidis (ATCC 12228) strains and SA, η , log P molecular descriptors.
Statistical parameters for the resulting regression equations: correlation coefficient
(R), determination coefficient (R2 ), adjusted determination coefficient for calibra-
tion (R2 ad j ), a standard error of estimate (SEE), determination coefficient of LOO
validation (Q2 )
Strain Equation R2 R2 adj. F p SEE Q2
log(1/MICS.aur.ATCC29213 ) =
S. aureus ATCC
0.0011(±0.003)SA+ 0.561 0.488 7.7 0.007 0.291 0.358
29213
16.60(±6.08)η − 67.51(±23.94)
S. aureus NCTC log(1/MICS.aur.NCTC4163 ) =

0.396 0.349 8.5 0.012 0.278 0.287
4163 0.60(±0.21) log P − 3.28(±0.61)
S. aureus ATCC 0.0012(±0.004)SA+ 0.487 0.402 5.7 0.018 0.358 0.082
25923
19.42(±6.21)η − 77.95(±23.94)
0.30 0.25 5.7 0.030 0.40 0.179
0.70(±0.30) log P − 3.46(±0.87)
S. aureus ATCC 0.0011(±0.003)SA+ 0.487 0.402 5.7 0.018 0.358 0.082
6538
17.06(±7.49)η − 69.43(±28.34)
0.290 0.236 5.3 0.040 0.360 0.183
0.60(±0.26) log P − 3.28(±0.77)
log(1/MICS.epi.ATCC12228 ) =
S. epidermidis 0.0011(±0.003)SA+ 0.552 0.477 7.9 0.009 0.297 0.192
ATCC 12228
18.26(±6.21)η − 73.54(±23.91)
log(1/MICS.epi.ATCC122228 ) =
0.330 0.278 6.4 0.030 0.348 0.151
0.60(±0.26) log P − 3.28(±0.77)

Fig. 19.5: Correlation of activity against S. aureus ATCC6538 and SA, η .

Fig. 19.6: Correlation of activity against S .aureus ATCC29213 and SA, η .


Fig. 19.7: Correlation of activity against S. epidermidis ATCC12228 and SA, η .
ized by the highest values of SA, Rf, α , V, SAg, but it does not show the highest
biological activity. Its average antimicrobial activity against all analyzed strains is
MIC = 22.4 [μ g/ml]. Group B contains compounds no. 12 and 13 incorporating in
the R1 position a phenyl ring with trifluoromethyl substituted in the meta or para
position, as well as compounds no. 11 and 6 with a phenyl ring substituted with an
atom of iodine or chlorine and methyl. Group C consists of six compounds no. 14,
5, 4, 9, 10 and 2. Compounds no. 5, 4, 9, 10 and 2 incorporate in the R1 position
a phenyl ring substituted in different positions with a single atom of bromine or
chlorine. The last group D contains compounds no. 7, 8, 3 and 1 with a phenyl ring
substituted in different positions with a single fluorine atom or not substituted at
all. Groups obtained by grouping compounds on the basis of relationships with de-
scriptors with determined linear dependencies do not coincide with groups obtained
by grouping compounds on the basis of their biological activity. Group D contains
compounds with the lowest antimicrobial activity from MIC = 40 [μ g/ml] to MIC
= 200 [μ g/ml].
19.4 Conclusions
The conducted statistical analysis enabled determination of linear relationships be-

tween antimicrobial activity against Staphylococcus aureus (NCTC 4163, ATCC
25923, ATCC 6538, ATCC 29213), Staphylococcus epidermidis ATCC 12228 bac-

Fig. 19.8: Dendrogram of compound grouping based on relationships with three

molecular descriptors SA, log P, η .
teria and two molecular descriptors SA, η . The obtained equations indicate a posi-
tive correlation between both descriptors and biological activity. The chemometric
analysis performed showed that compounds with a phenyl ring substituted in ortho,
meta and para positions with an atom of fluorine are characterized by the lowest
antimicrobial activity values and the lowest Rf, log P, SA descriptor values. The
conducted analyses indicate that the increase in activity against the studied strains
is closely related to the type and position of substituent in a phenyl ring. The para
position is preferred in this case.
Acknowledgements
This work was partially supported by the Ministry of Science and Higher Educa-
tion funding for statutory activities of young researchers of Faculty of Automatic
Control, Electronics and Computer Science (Silesian University of Technology).
This work was partially supported by the Ministry of Science and Higher Education
funding for statutory activities of researchers of Faculty of Biomedical Engineering
(Silesian University of Technology).
References
1. Leonetti, F., Favia, A., Rao, A., Aliano, R., Paluszcak, A., Hartmann, R.W., Carotti, A. Design,
synthesis, and 3D QSAR of novel potent and selective aromatase inhibitors. J. Med. Chem.,
30, 6792-803, (2004)
2. Pierre, F., Stefan. E., Nédellec, A.S., Chevrel, M.C., Regan, C.F., Siddiqui-Jain, A., Macalino,
D., Streiner, N., Drygin, D., Haddach,M., O’Brien, S.E., Anderes, K., Ryckman, D.M. 7-(4H-
1,2,4-Triazol-3-yl)benzo[c] [2,6]naphthyridines: a novel class of Pimkinase inhibitors with
potent cell antiproliferative activity. Bioorg. Med. Chem. Lett., 15, 6687-92, (2011)
3. Kumar, D., Narayanam, M.K., Chang, K.H., Shah. K. Synthesis of novel indolyl-1,2,4-
triazoles as potent and selective anticancer agents. Chem. Biol. Drug. Des., 77, 182- 188,
(2011)
4. Sumangala, V., Poojary, B., Chidananda, N., Arulmoli, T., Shenoy,S. Facile synthesis, cyto-
toxic and antimicrobial activity studies of a new group of 6-aryl-3-[4-(methylsulfonyl)benzyl]-
7H-[1,2,4]triazolo[ 3,4-b][1,3,4]thiadiazines. Eur. J. Med. Chem., 54, 59-64, (2012)
5. Ji, D., Lu, J., Lu, B., Xin, C., Mu, J., Li, J., Peng, C., Bao, X. Efficient synthesis and antimi-
crobial activity of some novel S-_-Dglucosidesof 5-aryl-1,2,4-triazole-3-thiones derivatives.
Bioorg. Med. Chem. Lett.„ 23, 1997-2000, (2013)
6. Koparir, M., Orek, C., Parlak, A.E., Söylemez, A., Koparir, P., Karatepe, M., Dastan, S.D.
Synthesis and biological activities of some novel aminomethyl derivatives of 4-substituted-5-
(2-thienyl)- 2,4-dihydro-3H-1,2,4-triazole-3-thiones. Eur. J. Med. Chem., 63, 340-346, (2013)
7. Prakash, O., Aneja, D.K., Hussain, K., Lohan, P., Ranjan, P., Arora, S., Sharma, C. Aneja,
K.R. Synthesis and biological evaluationof dihydroindeno and indeno [1,2-e] [1,2,4]triazolo
[3,4-b] [1,3,4]thiadiazines as antimicrobial agents. Eur. J. Med. Chem. 2011, Tom 46, strony
5065-5073.
8. Stefanska J., Stepien K., Bielenica A., Szulczyk D., Miroslaw B., Koziol A. E., Sanna G.,
Iuliano F., Madeddu S., Jozwiak M., Struga M. Antimicrobial and Anti-biofilm Activity of
Thiourea Derivatives Bearing 3-amino-1H-1,2,4-triazole Scaffold. Medicinal Chemistry., 12,
478-488, (2016)
9. Stefanska J., Szulczyk D., Koziol B. E., Miroslaw B., Kedzierska E., Fidecka S., Busonera
B., Sanna G., Giliberti G., La Colla P., Struga M. Disubstituted thiourea derivatives and their
activity on CNS: Synthesis and biological evaluation. European Journal of Medicinal Chem-
istry., 55, 205-213, (2012).
10. Hyper Chem. Hyper Chem (TM), Professional, Hypercube, Inc., 1115 NW 4th Street,
Gainesville, Florida 32601, U.S.A. (2012)
11. Stefanska J., Nowicka G., Struga M., Szulczyk D., Koziol A.E., Augustynowicz-Kopec E.,
Napiorkowska A., Bielenica A., Filipowski W., Filipowska A., Drzewiecka A., Giliberti G.,
Madeddu S., Boi S., La Colla P., Sanna G., Antimicrobial and anti-biofilm activity of thiourea
derivatives incorporating a 2-aminothiazole scaffold, Chemical and Pharmaceutical Bulletin
(Tokyo), 63, 225-36 (2015)
12. Bodor N., Gabanyi Z., Wong C., A new method for the estimation of partition coefficient, J.
Am. Chem. Soc.,111, 3783 (1989)
13. Filipowska A., Filipowski W., Tkacz E., Nowicka G., Struga M.,. Statistical analysis of
the impact of molecular descriptors on cytotoxicity of thiourea derivatives incorporating 2-
aminothiazole scaffold. Chemical and Pharmaceutical Bulletin., 64, 1196-1202, (2016)
14. Astela A., Biziukb M., Przyjaznyc A., NamiesÂt’nikb J., Chemometrics in monitoring spatial
and temporal variations in drinking water quality, Water Research, 40, 1706–1716, (2006)
15. StatSoft, Inc. (2014). STATISTICA (data analysis software system), version 12.
www.statsoft.com
16. Dehmer M., Varmuza K., Bonchev D., Emmert-Streib F., Statistical Modeling of Molecular
Descriptors in QSAR/QSPR, Wiley-Blackwell, (2012)
17. Khaledian S., Saaidpour S. Quantitative Structure-property Relationship Modelling of Dis-

tribution Coefficients (logD7.4) of Diverse Drug by Sub-structural Molecular Fragments.
Method Orient. J. Chem., 31, 1969-1976, (2015)
18. Filipowska A., Filipowski W., Tkacz E.,. Study of structure-cytotoxicity relationships of
thiourea derivatives containing the 2-aminothiazole moiety, Innovations in biomedical engi-
neering, Advances in Intelligent Systems and Computing. Innovations in Biomedical Engi-
neering: Springer , 526, 276-285, (2017)
19. Carbó-Dorca R., Robert D., Amat L., Girones X., Besalu E. Molecular Quantum Similarity
in QSAR and Drug Design, Lecture notes of chemistry 73, Springer-Verlag Berlin Heidelberg
(2000)
20. Fukui K., Yonezawa T., Shingu H. A Molecular Orbital Theory of Reactivity in Aromatic
Hydrocarbons. J. Chem. Phys., 20, 722 âĂŞ 725, (1952)
Chapter 20
The Face Tracking System for Rehabilitation
Robotics Applications
Paweł Raif1 , Ewaryst Tkacz1
Abstract. The paper presents the working model of the face tracking system. The
proposed solution may be used as one of the parts of the rehabilitation or assistive
robotic system and serve as the robotic vision subsystem or as the module con-
trolling robotic arm. It is a low-cost design, it is based on open source hardware and
software components. As a hardware base the Raspberry Pi computer was used. The
machine vision software is based on Python programming language and OpenCV
computer vision library.
Keywords: machine vision, rehabilitation robotics, assistive robotics, human com-

puter interaction, face recognition, object tracking, OpenCV, Raspberry Pi
20.1 Introduction
The goal of rehabilitation, is to enable a person to regain the maximal possible

level of independence or productivity. Robots can help in therapy of individuals
suffering from many kinds of motor impairments, for example due to illness, stroke
or accidents. Rehabilitation and assistive robotics are very active research topics.
The motivation for the research in these fields comes from different places. One
of them is stroke rehabilitation and stroke recovery [1]. Assistive robotics, and in
general assistive technologies, are becoming more important due to ageing society
[2].
Interactive robots equipped with face tracking systems can be used also in autism
therapy where they serve as social mediators for promoting and teaching communi-
cation skills in autistic children [3]. Tracking systems are also important in situations
Department of Biosensors and Processing of Biomedical Signals,

Biomedical Engineering Faculty at the Silesian University of Technology,
Zabrze, Poland

https://doi.org/10.1007/978-3-319-70063-2_20
186 Paweá Raif, Ewaryst Tkacz
when social human-robot interaction is important, when robots should be socially

interactive [4].
In almost all of the mentioned situations the computer vision systems in general
and computer vision based object tracking systems in particular, can be applied.
The face or eyes tracking systems are usually vital parts of the rehabilitation and
assistive robots [5].
The face tracking system presented in this paper can serve as a part of rehabili-
tation or assistive robotic system. It can be used for example as a robot vision sub-
system, or as a module for robotic arm control (for example arm carrying cameras,
sensors, diagnostic or assistive devices). It can be also used as a part of autonomous
cognitive robotic system based on motivated learning method [6][7].
The proposed solution is implemented using only open source hardware and soft-
ware technologies. We have used Raspberry Pi computer with Raspbian operating
system. We have used software technologies based on Python and C++ program-
ming languages (like OpenCV or SimpleCV computer vision libraries).
In the subsequent sections of the document we have presented as follows: in
Methods section: the hardware and software elements, the structure of the system
(connections between its parts) and block and control flow diagrams. Our observa-
tions and remarks about the performance of the face tracking system are presented
in sections: Results, Conclusions and Future work.
20.2 Methods
20.2.1 Hardware
The proposed system is based on open-source computer hardware Raspberry Pi [8].

The Raspberry Pi is a credit card-sized single-board computer. It is developed by
the Raspberry Pi Foundation.
The used Raspberry Pi (RPI) B model is based on the Broadcom BCM2835
system on a chip (SoC), which includes an ARM1176JZF-S 700 MHz processor,
VideoCore IVGPU, and 512 Mb of RAM. The RPI system has also Secure Digital
(SD) memory card connector for operating system and user programs. The Rasp-
berry Pi computer can use several dedicated operating systems (mostly linux dis-
tributions, for example Raspbian, OSMC, OPENELEC, XBMC, but there are also
available a non linux distributions, for example RISC OS). In our system and exper-
iments we used standard Raspbian operating system [9].
In our system we have used Raspberry Pi computer, Raspberry PI 5MP Camera
Board Module (in initial tests we have used simple Logitech web camera) and two
TowerPro SG90 9g mini servos to control camera movements.
20 Rehabilitation Robotics 187
20.2.2 Software
In proposed system we have used open source scientific software technologies based
on Python programming language [10][11]. This includes: Numpy, Scipy, Pygame.
As a machine vision and image analysis software we have used OpenCV library
with SimpleCV interface.
The OpenCV (Open Source Computer Vision) [12] library is free for both aca-
demic and commercial use, it is released under a BSD license. The OpenCV is
written in C++ (hence its primary interface is in C++), but it has full interfaces in
Python, Java and Matlab. It supports several operating systems: Windows, Linux,
Mac OS and Android. In our experiments we have used its distribution for Raspbian
linux (and for Ubuntu Linux in some initial tests).
The SimpleCV (Simple Computer Vision) [13][14] is an open source framework
for building computer vision applications. It includes several high-powered com-
puter vision libraries âĂŞ among them OpenCV. It uses Python for scripting (there-
fore it is possible to either run scripts or use interactive shell) so it is very useful for
quick prototyping.
The remaining libraries: NumPy (numerical Python) [15], SciPy (scientific Python)
[16] provide efficient numerical computations, scientific computing and visualiza-
tion respectively. The Pygame library provides both window and screen manage-
ment functions [17].
20.2.3 Connections
The block diagram of the proposed system is shown in the figure 20.1.
Fig. 20.1: Block diagram

In its most basic form face tracking system consists of Raspberry Pi computer
with dedicated camera module (Raspberry PI 5MP Camera Board Module) con-
nected via CSI port (Camera Serial Interface) and camera mount.
In the presented prototype the camera body is connected to the two axis camera
mount constructed with two analog micro servomachanisms TowerPro SG92R 9g.
Hence each servo operating angle is almost 180 degrees (Âś 90 degrees of angu-
lar freedom from neutral position), we can assume that the camera field of view is
almost hemispherical.
20.2.4 Camera Control
The camera position is controlled by two micro servos TowerPro SG92R 9g, one for
x-axis movement and the other for y-axis movement.
The typical servo is connected through a three wire connection: +, − and S (sig-
nal). The servo operation is controlled by PWM (Pulse Width Modulation) signal
which is a series of variable width pulses. The pulse is repeated every 20 ms but it is
not critical. Critical is the pulse width because it determines the position of the arm
of the servo. Typically the pulse width is between 0.5 ms and 2.5 ms.
In our tracking system micro servos are controlled by Raspberry Pi GPIO (Gen-
eral Purpose Input and Output) 26 pin port. The GPIO port can control only one
servo. In order to avoid this obstacle one can use ServoBlaster driver [18] which
allows to generate PWM signals from any of Raspberry Pi’s GPIO pins. The GPIO
port can provide 5V. The connection between Raspberry Pi’s GPIO port and ser-
vomechanisms is shown in the figure 20.2.
The servo x, controlling x axis rotation is connected to GPIO 23 pin, and servo
y, controlling y axis rotation, to GPIO 24 pin. Both of the micro servomechanisms
are controlled using Python programming language.
A Python module dedicated to control Raspberry Pi GPIO channels is RPi.GPIO
(or its extension: RPIO module which has which has built-in support for software
PWM so doesn’t need the ServoBlaster driver).
20.2.5 Vision analysis and face tracing task
The control flow diagram for the proposed tracking system is depicted in the figure
20.3.
After the start of the system both servomechanism are set to neutral position and
then real time visual face recognition starts to work. When it detect specified object
in the camera‚s field of view, coordinates of the object are send to functione con-
trolling camera mount servos. Being provided with the coordinates of the position
of the detected object program checks if the object is positioned in the center of the
camera view (or other desired region). If not, tracking system calculates the rotation
Fig. 20.2: Servos connected to GPIO port.
Fig. 20.3: Control flow diagram

angles for both x and y axis. Compute servo’s arm rotations, start the servos and
move camera to proper direction.
The object detection task is performed by functions of the OpenCV library [19].
There are freely available effective classifiers for objects like face, nose or eyes.
They are implemented in the OpenCV library. One of them is the Haar cascade
classifier [20], presented in [21][22]. The Haar cascade classifier can be trained in
order to detect any visual object, but in our project we have used only predefined
patterns like face, nose or eye.
The photo of the prototype of the tracking system camera mount is shown in the
figure 20.4.
Fig. 20.4: The prototype: Raspberry Pi Camera with two axis camera mount con-
structed with two micro servos.
20.3 Results
The OpenCV computer vision library is very efficient and useful tool for such appli-
cations like visual object recognition or object tracking. Supported with Simple CV
library it allows quick prototyping for any kind of machine vision projects (applied
to robotics, biomedical engineering, real time image and video analysis).
The implemented face tracking system works very well as a prototype, but in the
future real applications some efficiency improvements would be necessary.
20.4 Conclusions and future work
The proposed object tracking system after efficiency improvements can be used as a
part of rehabilitation or assistive robotic system (for example as a vision subsystem
or module for robotic arm control).
First and foremost, however, the algorithms and their implementation in OpenCV
library are very efficient, performed task (real time face recognition) is too demand-
ing for computers like Raspberry Pi. In order to improve the efficiency of the whole
system we should experiment with other, more efficient hardware, for example:
Raspberry Pi overclocked to 800 MHz, Raspberry Pi B+, Banana Pi, Beagle Board
Black.
In case of using the proposed tracking system as a part of vision subsystem, it
probably would be better to use the brushless camera mount gimbal instead of the
micro servomechanisms. In that case the tracking mechanism would work more
quietly and more quickly. It would also provide more steady image for the analysis.
When used as a vision subsystem, it could be provided with two cameras on the
same mount: one with wide-angle lens (its task would be to find the specified object
and center the camera on it), and the other with long-focus lens (its task would be
to examine selected object more precisely).
In case of using proposed tracking system as a subsystem for controlling robotic
arm the whole two axis servos based mount could be 3D printed. Therefore it would
be more mechanically stable and could be used as a base for robotic arm carrying
for example specific actuators or diagnostic sensors.
References
1. H. Zhou, H. Hu: Human motion tracking for rehabilitation - A survey, Biomedical Signal
Processing and Control, Volume 3, Issue 1, pp. 1âĂŞ18, January 2008.
2. C. Rougier, J. Meunier, A. St-Arnaud, J. Rousseau: Monocular 3D Head Tracking to Detect
Falls of Elderly People. Proceedings of the 28th IEEE EMBS Annual International Conference
New York City, USA, Aug 30-Sept 3, 2006.
3. L. Boccanfuso, J. M. O’Kane: Adaptive Robot Design with Hand and Face Tracking for Use in
Autism Therapy. Social Robotics, Second International Conference on Social Robotics, ICSR
2010, Singapore, November 23-24, 2010, pp. 265-274, DOI: 10.1007/978-3-642-17248-9_28.
4. T. Fong, I. Nourbakhsh, K. Dautenhahn: A survey of socially interactive robots. Robotics and
Autonomous Systems, Volume 42, Issues 3âĂŞ4, pp. 143âĂŞ166, March 2003.
5. P. Jia, H. H. Hu, T. Lu, K. Yuan: Head gesture recognition for hands free control of an in-
telligent wheelchair. Industrial Robot: An International Journal, Vol. 34 Iss: 1, pp.60 âĂŞ 68,
2007.
6. P. Raif, J.A. Starzyk: Motivated learning in autonomous systems. The 2011 International Joint
Conference on Neural Networks (IJCNN), pp. 603-610, 2011.
7. J.A. Starzyk, J.T. Graham, P. Raif, A.H. Tan: Motivated Learning for Autonomous Robots
Development. Cognitive Science Research, 14, 1, 2011.
8. RaspberryPi, https://www.raspberrypi.org/
9. Raspbian, http://www.raspbian.org/
10. Python, http://www.python.org/
11. J.E. Solem: Programming Computer Vision with Python: Tools and algorithms for analyzing
images. O’Reilly Media 2012.
12. OpenCV, http://opencv.org/
13. SimpleCV, http://simplecv.sourceforge.net/, http://simplecv.org/
14. K. Demaagd, A. Oliver, N. Oostendorp, K. Scott: Practical Computer Vision with SimpleCV:
The Simple Way to Make Technology See. O’Reilly Media 2012.
15. NumPy, http://numpy.scipy.org/
16. SciPy, http://www.scipy.org/
17. Pygame, http://pygame.org/
18. ServoBlaster, https://github.com/richardghirst/PiBits/tree/master/ServoBlaster
19. J. Howse: OpenCV Computer Vision with Python. CreateSpace Independent Publishing Plat-
form 2015.
20. G. Bradski, A. Kaehler: Learning OpenCV: Computer Vision with the OpenCV Library.
O’Reilly Media 2008.
21. P. Viola, M. Jones: Rapid object detection using a boosted cascade of simple features. In: Proc.
IEEE Computer Society Conference on Computer Vision and Pattern Recognition. vol. 1, pp.
511–518, 2001.
22. P. Viola, M. Jones: Robust Real-time Object Detection. International Journal of Computer
Vision 57(2), pp. 137-154, 2004.
Chapter 21
The higher-order spectra as a tool for the
identification of patients diagnosed with various
cardiac diseases
Zbigniew Budzianowski, Ewaryst Tkacz, Wojciech Oleksy, Małgorzata Garbacik
Abstract. This article explores the possibility of using the higher-order spectra to
identify different types of diseases. In order to assess the effectiveness of such tool
the HRV (Heart Rate Variability) recordings obtained from patients suffering from
three different cardiac problems are listed and compared to the results recorded for
healthy subjects. Each set of HRV signals is processed with bispectral and bicoher-
ent analysis. In both cases three statistical parameters are observed. For each type of
the investigated analysis the parameters under examination differ enough to allow
clear distinction of the specific cardiac disease. The obtained results show useful-
ness of higher-order spectra as a tool for differentiation between specific diseases.
Authors believe that further work would greatly improve potential of the described
tool, allowing to identify number of different diseases or even stage of the illness or
progress in the rehabilitation process.
Keywords: Heart rate variability, higher-order statistics, signal processing
21.1 Introduction
The aim of this work was to answer the question whether higher-order spectral anal-
ysis , based on the records of heart rate variability, could be a valuable diagnos-
tic tool that allows the identification of specific diseases. To achieve this the HRV
registrations taken from four different groups of subjects were compared. The first
group consisted of healthy individuals and served as a reference one. Three other
groups consisted of people burdened with different cardiac diseases: arrhythmia,
tachyarrhythmia and congestive heart failure. For each of these groups bispectral
and bicoherent analyses were performed. The results are summarized below.
Silesian University of Technology, 41-800 Zabrze, Poland,

e-mail: zbigniew.budzianowski@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_21
194 Zbigniew Budzianowski et al.
21.2 Material and Methods
21.2.1 A description of the data used
All the used data were collected from the sets of signals stored in an online database
PhysioNet [1]. The main source of comparison have become records collected under
the file name Normal Sinus Rythm RR Interval Database (nsr2db) [1]. This database
includes beat annotation for long-term ECG recordings of subjects in normal sinus
rhythm. For comparative purposes three other sets were also used:
• MIT-BIH Arrhythmia Database (mitdb) - HRV signals database of patients di-

agnosed with arrhythmia (A) [1].
• CU Ventricular Tachyarrhythmia Database (cudb) - HRV signals database of
patients diagnosed with tachyarrhythmia (TA) [1].
• Congestive Heart Failure RR Interval Database (chf2db) - HRV signals database
of patients diagnosed with congestive heart failure (CHF) [1].
21.2.2 Signal processing
In order to eliminate artifacts simple filtering mechanism was applied. In case of

each long-term recording the average length of RR interval was calculated. Next,
the length of each RR interval in this recording was compared with the previously
computed average. If the length of interval was shorter than 80% or longer than
120% of the average, such interval was rejected. It was assumed that all the in-
tervals that remained after filtration were NN ones. In case of each bispectral and
bicoherent analysis authors used the first 256 registered NN intervals, dividing them
into 8 equal segments (each of 32 intervals) [2]. The analysis was performed using
Matlab’s HOSA Toolbox. [2] Each analysis was performed using the direct method
[3], [4].
21.3 Calculation
Frequency analysis allows the separation of the individual components of the spec-
trum. This type of study allows to detect cyclicality in change of the NN intervals
length [5], [6], [7]. The aim of the analysis in the frequency domain is to decom-
pose the total variability of NN intervals into individual frequency components. The
result is a plot of the power spectrum as a function of frequency [8]. To evaluate
the total spectra power of NN intervals variability the following parameters are used
[6]:
21 HOSA as a ident. tool 195
• ULF - ultra low frequency component (under 0,0033Hz). The value expressed
in ms2
• VLF - very low frequency component (between 0.0033 to 0.04 Hz). The value
expressed in ms2.
• LF - low frequency component (from 0.04 to 0.15 Hz). The value expressed in
ms2
• HF - high frequency components (from 0.15 to 0.4 Hz). The value expressed in
ms2
During the work it was decided not to treat ULF band as a separate one. Instead, it
is assumed that VLF band covers a range from 0 to 0.04 Hz.
Table 1 shows final division into six regions of analysis based on combinations of
four sub-bands:
Table 21.1: Regions of bispectral and bicoherent analyses
Region Range Band f1 [Hz] Band f2 [Hz]

1 VLF-VLF 0 - 0.04 0 - 0.04
2 LF-VLF 0.04 - 0.15 0 - 0.04
3 LF-LF 0.04 - 0.15 0.04 - 0.15
4 HF-VLF 0.15 - 0.4 0 - 0.04
5 HF-LF 0.15 - 0.4 0.04 - 0.15
6 HF-HF 0.15 - 0.4 0.15 - 0.4
Due to the symmetry of the bispectrum [9], [10], regions 1, 3 and 6 are limited
by a diagonal of coordinates 0, 0, and 0.5, 0.5. The final distribution of the analyzed
regions is illustrated below.
In each of the six analyzed regions the maximum and average values of bis-
pectrum are calculated along with its variance. The results are presented in tabular
form.
21.4 Results
The charts below show the values obtained during bispectral and bicoherent analyses
of HRV records. As mentioned before, there were four groups. Three of them were
patients suffering from different cardiac diseases. The last one was the reference
(healthy ones). Objective of the study was to verify whether it’s possible to identify
specific diseases based on the data obtained from HRV signal.
Fig. 21.1: Distribution of the analyzed regions
21.4.1 Bispectral analyses
21.4.1.1 Comparison of the maximum values of the bispectra
In order to improve the readability of the chart the logarithmic scale was used.
Table 21.2: Comparison of the maximum values of the bispectra
Band Ref. group A TA CHF

VLF-VLF 1,24E+04 3,34E+05 4,96E+03 2,08E+04
LF-VLF 1,49E+04 1,18E+05 2,91E+03 8,00E+03
LF-LF 9,21E+03 2,31E+04 3,92E+03 8,08E+03
HF-VLF 1,86E+03 1,98E+04 9,55E+03 2,78E+03
HF-LF 1,84E+03 1,08E+04 6,73E+03 2,55E+03
HF-HF 637,07 4,22E+03 1,52E+04 3,69E+03
Observations:
• Observing the maximum value of bispectrum in different frequency ranges, one
can see that in each band, other than HF-HF one, results from patients suffering
from arrhythmia exceed the results calculated for other groups. This difference
is particularly evident in the case of the lowest frequencies (VLF-VLF and LF-
VLF bands).
• In the VLF- VLF band maximum value of bispectrum calculated for the group
suffering from arrhythmia is 16 times higher than the second highest result
(CHF)
Fig. 21.2: Comparison of the maximum values of the bispectra
21.4.1.2 Comparison of the average values of the bispectra
Table 21.3: Comparison of the average values of the bispectra

VLF-VLF 5,01E+03 1,28E+05 1,40E+03 7,63E+03
LF-VLF 1,96E+03 1,25E+04 610,12 944,5
LF-LF 607,99 260,04 346,69 412,23
HF-VLF 237,83 3,62E+03 984,52 398,75
HF-LF 84,60 191,76 395,54 424,46
HF-HF 19,37 190,38 977,76 335,01
Observations:
• Observing the average values of bispektra in different frequency ranges , one
can see that the in case of VLF-VLF , LF-VLF and HF-VLF bands results of the
group suffering from arrhythmia exceed the results calculated for other groups.
• This difference is particularly evident in the case of VLF - VLF band, where
average value of bispectrum calculated for the group suffering from arrhythmia
is over 16 times higher than the second highest result (CHF)
21.4.1.3 Comparison of the variance of the bispectra
Observations:
Fig. 21.3: Comparison of the average values of the bispectra
Table 21.4: Comparison of the variances of the bispectra

VLF-VLF 1,70E+07 1,26E+10 4,61E+06 6,24E+07
LF-VLF 2,09E+07 6,14E+08 8,19E+05 5,76E+06
LF-LF 3,60E+06 1,65E+07 2,16E+06 1,80E+06
HF-VLF 4,36E+05 3,16E+07 6,51E+06 1,50E+06
HF-LF 2,80E+05 4,22E+06 4,73E+06 5,59E+05
HF-HF 5,06E+04 1,61E+06 7,85E+06 5,88E+05
Fig. 21.4: Comparison of the variances of the bispectra
• Comparing the variances in the different bands, once again one can observe that
the results calculated for group suffering from arrhythmia stand out.
• The biggest difference can be spotted in the VLF-VLF band. However, this time
it’s up to 800 times higher than in the case of the second highest result (CHF)
21.4.2 Bicohrent analyses
21.4.2.1 Comparison of the maximum values of the bicoherence
The results are given below.
Table 21.5: Comparison of the maximum values of the bicoherence

VLF-VLF 0,1519 0,068 0,2377 0,2546
LF-VLF 0,4163 0,2259 2,6433 0,1613
LF-LF 0,7494 0,5448 6,7156 0,2196
HF-VLF 0,6368 0,4355 2,5818 0,3043
HF-LF 0,2708 1,2562 5,8711 0,2638
HF-HF 0,3638 1,7753 6,2143 0,6147
Fig. 21.5: Comparison of the maximum values of the bicoherence
Observations:
• In each band , except the VLF-VLF one, the maximum value of bicoherence
obtained for a group suffering from tachycardia significantly exceeds other reg-
istrations.
• The biggest difference was recorded in the LF-LF band, where the result calcu-
lated for patients with tachycardia was almost 9 times higher than the second
highest (reference group).
21.4.2.2 Comparison of the average values of the bicoherence
The results are given below.
Table 21.6: Comparison of the average values of the bicoherence

VLF-VLF 0,0183 0,0111 0,0178 0,1194
LF-VLF 0,0997 0,0904 1,666 0,038
LF-LF 0,1313 0,1489 4,0585 0,0649
HF-VLF 0,1012 0,1968 2,1302 0,1119
HF-LF 0,0673 0,4638 2,8375 0,0703
HF-HF 0,592 0,7818 3,1083 0,0855
Fig. 21.6: Comparison of the average values of the bicoherence
Observations:
• In each band , except the VLF-VLF one, the maximum value of bicoherence
obtained for a group suffering from tachycardia significantly exceeds other reg-
istrations.
• The biggest difference was recorded in the LF-LF band, where the result calcu-
lated for patients with tachycardia was almost 9 times higher than the second
highest (reference group).
21.4.2.3 Comparison of the variances of the bicoherence
Table 21.7: Comparison of the variances of the bicoherence

VLF-VLF 0,001 9,75E-05 0,0011 0,0076
LF-VLF 0,0077 0,0031 0,6495 0,0012
LF-LF 0,0176 0,0115 1,4728 0,0024
HF-VLF 0,0124 0,0117 0,0488 0,003
HF-LF 0,0041 0,0543 1,5882 0,0032
HF-HF 0,0054 0,1101 1,6882 0,0074
Fig. 21.7: Comparison of the variances of the bicoherence
Observations:
• Just like in the case of a comparison of the maximum values of bicoherence,
in each band, except the VLF-VLF one, the results of the group suffering from
tachycardia clearly stand up .
• The highest difference can be observed in the LF-LF band. The average value
of bicoherence is more than 26 times higher than the second highest result (ar-
rhythmia).
21.4.3 Discussion
21.4.3.1 Bispectral analysis - summary
Comparison of the results of the reference group and patients suffering from ar-
rhythmia, tachycardia and CHF showed the usefulness of the bispectral analysis as
a tool for differentiation between specific diseases. For each of the analyzed parame-
ters one can easily distinguish results registered for patients suffering from arrhyth-
mia. The differences between arrhythmia and other cases were so significant that
high effectiveness of such identification can be assumed. Particularly in the case of
comparing the variations in VLF-VLF band. The variation obtained for the patients
suffering from arrhythmia was over 800 times higher than in the any other observed
case.
21.4.3.2 Bicoherent analysis - summary
Comparison of the results of the reference group and patients suffering from arrhyth-
mia, tachycardia and CHF showed the usefulness of the bicoherent analysis as a tool
for differentiation between specific diseases. For each of the analyzed parameters a
significant alternation of the results registered for patients suffering from tachycar-
dia can be observed. The biggest differences occurred in the LF-LF band. They were
so significant that high efficiency of such identification can be assumed. Particularly
in the case of comparing the variations in the above-mentioned frequency range,
where the result obtained for the tachycardia was over 80 times greater than other
ones.
21.4.4 Conclusions
As the result of the work the analysis of the suitability of the higher-order spectra
as a tool for the identification of patients diagnosed with various cardiac diseases
was presented. To simplify, only three parameters were taken under consideration
(variance, maximum and average value). Each of them can be used to identify reg-
istrations taken from healthy individuals and patients suffering from arrhythmia or
tachycardia. However, in each case the most robust analysis was the one based on
comparison of the variances. Only for a third group, that is the patients diagnosed
with CHF the results were not conclusive. Therefore, it seems that the key to im-
proving the identification mechanism is to extend the list of investigated parameters.
In the present work this step was abandoned due to the fact that the efficiency of the
described analysis has already been confirmed for the groups with tachycardia and
arrhythmia.
21.5 List of Abbreviations
• A - Arrhythmia
• CHF - Congestive Heart Failure
• ECG - Electrocardiogram
• HF - High frequency
• HOSA - Higher Order Spectral Analysis
• HRV - Heart Rate Variability
• LF - Low frequency
• NN - Normal-to-normal intervals
• RR - Time intervals between consecutive QRS complexes
• TA - Tachyarrhythmia
• ULF - Ultra low frequency
• VLF - Very low frequency
References
1. PhysioNet [Online], available: http://physionet.org/, last access: March 2015.

2. A. Swami, J. Mendel i L. Chrysostomos Higher-Order Spectral Analysis Toolbox: For Use with
MATLAB 1993: The Mathwork.
3. A. Goshvarpour i A. Goshvarpour Comparison of higher order spectra in heart rate signals
during two techniques of meditation: Chi and Kundalini meditation 2013: Cognitive Neurody-
namics.
4. C. Chua Analysis of Cardiac and Epileptic Signals Using Higher Order Spectra 2010: PhD
Thesis, Queensland: Queensland University of Technology.
5. R. Gałaska
˛ Analiza fraktalna zmiennoÅŻci rytmu zatokowego u pacjentow z uposledzona
funkcja lewej komory miesnia serwcowego 2006: PhD Thesis, Gdansk: Akademia Medyczna
w Gdansku.
6. T. Krauze, P. Guzik i H. Wysocki, Zmiennosc rytmu serca: aspekty techniczne 2001: Nowiny
lekarskie.
7. P. Mazur, R. Pfitzner i P. Matusiki, Analiza parametrow czestotliwosciowych zmiennosci rytmu
serca po pomostowaniu aortalno-wiencowym 2011: Folia Cardiologica.
8. M. Klopocka, J. Budzynski, R. Bujak, M. Swiatkowski, W. Sinkiewicz i M. Ziolkowski,
Analiza parametrow czestotliwosciowych zmiennosci rytmu serca po pomostowaniu aortalno-
wiencowym 2011: Folia Cardiologica.
9. I. Jouny i R. Moses, The Bispectrum of Complex Signals: Definitions and Properties 1992:
IEEE Transactions on Signal Processing.
10. S. Saliu, A. Birand i G. Kudaiberdieva, Bispectral Analysis of Heart Rate Variability Signal
2002.
Chapter 22
The prototype of wearable sensors system for
supervision of patient rehabilitation using
artificial intelligence methods
Eliasz Kańtoch1 , Dominik Grochala1 Marcin Kajor1 , and Dariusz Kucharski1
Abstract. In this paper, we investigate a wearable multi-sensor system for telemed-

ical supervision of patient rehabilitation in home conditions. Our approach is based
on a set of wearable sensors connected with a master digital acquisition module.
Proposed platform is capable of recording a single-lead ECG and acceleration sig-
nals which are used for determining the patient’s activity form. During research we
examined the following activities of daily living: sitting (idle), walking and squat-
ting. A set of machine learning methods were used for classification of mentioned
activities as rehabilitation exercises. Our methodology showed high success rates
(above 92%) for detecting selected activities and it showed great potential to be
used for patient’s activity recognition.
Keywords: wearable system, body sensor network, biomedical sensors, telemedicine,

patient rehabilitation, activity classification
22.1 Introduction
An automatic recognition of human activity is a novel approach for supporting phys-

iotherapist in the examination of patients’ rehabilitation progress outside the med-
ical facilities[1]. Constant development of technology and miniaturization of elec-
tronics caused rapid progress of wearable systems which are changing the face of
modern healthcare[2]. Currently there are many systems developed to investigate
a health status and analyze human physiological signals. Recently the impact has
been put on incorporating the Bluetooth standard into medical devices [3]It was
also stated that telemedical system with online user interface and mobile comput-
ing unit is the solution, which introduces vital functionalities desired for patients
and physicians[4]. Nowadays, recovery level of patient’s motoric system is mostly
AGH University of Science and Technology, 30 Mickiewicza Av., 30-059 Kraków, Poland

https://doi.org/10.1007/978-3-319-70063-2_22
206 Eliasz Kańtoch, Dominik Grochala Marcin Kajor, and Dariusz Kucharski
assessed manually in hospitals. Typical physical therapy includes patient’s exam-

ination followed by manual exercises, which usually should be repeated in home
conditions to obtain the desired efficacy[5]. The modern approach takes advantage
of wearable devices such as step counters and smartwatches which can track number
of daily steps or the distance covered. The vast majority of novel solutions is based
on the inertial sensors. Huang et. al stated that acceleration data is more effective
than gyroscope data in terms of movement analysis. Promising results are obtained
in systems combining these sensors with software providing feedback and therefore
stimulating accurate patient movement. Nevertheless, differentiation between daily
activities and rehabilitation exercises is still extremely challenging and therefore it
demands further research. This ambiguity is mainly caused by lack of movements
synchronicity and alternation of different repetitions of the same activity being per-
formed by the same person. The factors such as between-individual differences and
degrees of freedom also introduce significant uncertainty[5]. Concerning the men-
tioned difficulties, the artificial intelligence is a promising approach because it pro-
vides generalization of the classification algorithm. Nevertheless, machine learning
methods usually give the best results in conjunction with representative features
vectors of analyzed signals. Recent research state variety of approaches concern-
ing patient’s movement detection for supporting rehabilitation in home conditions.
Huang et. al used the Wireless 9DoF IMU sensors to examine seven different lower
limb exercises of 69 subjects. The data was transmitted to PC via Bluetooth with
the use of Multi Shimmer Sync software. After preprocessing the signals were seg-
mented with the Template Matching Segmentation algorithm. Then, the statistical
parameters, information theory features and time-frequency coefficients were ex-
tracted as feature vectors of signals. Variety of algorithms, including Decision Tree,
Multilayer Perceptron Neural Network and Support Vector Machine were used to
classify the incorrect and regular exercise movement [5]. Lu et. al took advantage
of binary pyroelectric sensors and wireless transmission module to examine paces
and gaits of subjects classified as injured or healthy. The rehabilitation treadmill was
used to expose the movement kinematics and prevent patients from falling. The sen-
sors were placed nearby the platform to capture the human gait geometry and pace.
An Autocorrelation Function was used to extract periodicity. Gait features, in turn,
were extracted by probabilistic models. The experiments proved the effectiveness
of proposed method which is a promising approach in supporting patients’ rehabil-
itation process[6].In the last few years robot-assistive devices have been playing an
important role in the modern rehabilitation scenarios. Saracino et. al. used the bionic
system MOTORE++, integrated with a set of inertial sensors and EMG signals to
design therapeutic procedure for upper-limb rehabilitation after stroke. In this work
the fusion of information form haptic robotic arm and wearable sensors was perform
to provide an assessment of motoric disabilities of patients. A set of parameters was
analyzed to successfully classify the injured and healthy subjects[7]. A similar ap-
proach, adapted in healthcare and detection of emergency situations gives promising
results[8] Pierleoni et al. confirmed high usefulness of systems that measure accel-
eration and barometric pressure for detecting events in terms of physical activity of
patients. An approach which is commonly adapted in current works involves com-
22 Wearable Sensors System 207
mercially available devices and sensors to record signals for further analysis[9].This
method certainly offers proven and reliable data acquisition. On the other hand im-
plementation of self-designed prototypes enables to personalize the measurement
chain and provides case-suited solution. This also broadens horizons and drives the
development of new devices which may be introduced in health care systems. In
this paper, we face the challenge of designing and prototyping a system based on
wearable sensors for supervision of rehabilitation process. A major benefit of this
approach is the possibility of continuously monitoring patient movement without
limitation of comfort in performing daily activities and rehabilitation exercises. In
our research, machine learning algorithms: Softmax Regression (SR), Multilayer
Perceptron (MLP), Support Vector Machine (SVM) and Random Forest (RF) were
used for classification of following activities: idle, walking and squatting in terms of
supervising the patient’s rehabilitation process. We chose squatting as rehabilitation
exercise, since no other lifting movement without barbell places as much stress and
strain on musculoskeletal system for stimulating the lower limb recovery[10].
22.2 Architecture of the developed prototype
22.2.1 System hardware
In this paper we propose multi-sensor measurement solution for rehabilitation ac-

tivity monitoring. The system was designed to simultaneously record signals from
multiple sensors. The solution is mainly based on accelerometers and ECG inte-
grated circuit. Apart from that, the device can also provide additional parameters,
such as temperature, barometric pressure, respiration rate and light intensity. The
data is stored on the microSD card and can be conveniently transfer to the personal
computer. Currently the development is focused on miniaturization and integration
of the device with smart textiles and rehabilitation equipment. The system architec-
ture was shown in the Figure22.1.
22.2.2 Multi-sensor acquisition unit
The central acquisition module is based on 8-bit AVR microcontroller ATmega328p

which provides sufficient computational performance while enabling application
simplicity. The designed prototype is equipped with MPU-6050, an advanced MEMS
unit which combines the 3-axis accelerometer with 3-axis gyroscope. This inte-
grated circuit is dedicated to precise motion analysis and utilizes the 16-bit A/D
converters providing the dynamic range from 2 to 16g. The communication between
the MPU-and microcontroller is established with I2C bus which facilitate further
expansion of the board. The measurement of barometric pressure is realized with
Fig. 22.1: Architecture of the system.
BMP-085 chip which additionally supports calculation of the above-the-sea-level

altitude. For temperature measurement the DS18b20 digital thermometer was im-
plemented using 1-Wire bus for data transfer. This device features 12-bit resolution
with ±0.5°C accuracy. Light intensity monitoring functionality is based on photore-
sistor voltage divider connected to the microcontroller’s ADC (in 10-bit resolution
mode). The designed device is powered by Li-Poly battery and 3.7V input voltage
is transformed to 3.3V standard supply level by DC/DC converters. This approach
ensures high energy efficiency. The average battery life reaches approximately 24
hours, which is extremely vital in terms of long-time patient monitoring.Fig. 22.1.
22.2.3 ECG acquisition
Additionally, the conducted experiments involved measurement of heart electrical

activity (ECG). The biopotential acquisition module was based on integrated ana-
log front-end circuit ADS1292R. It allowed to obtain one-lead ECG and respiration
signals recorded by 2 independent 24-bit channels with 500Hz sampling rate. In this
project the development kit ADS1x9xECG-FE was used for acquisition. Therefore,
the application of mentioned chip was simplified. It also ensured a stable operation
during experiment. The LabVIEW software(Fig.22.2)was used for acquisition and
initial signal processing (Fig. 2.). The implemented IIR filter ensured DC attenua-
tion at 22dB level. In order to eliminate mains distortion we used pre-implemented
band-stop and band-pass filtration (50 Hz and 40/150 Hz) algorithms. During mea-
surement, subjects were connected to the device by standard 4-wire topology (R, L,
M, N) and Ag/AgCl electrodes. Considering a safety issues, the laptop used for con-
ducting experiments was powered by battery and electrically separated from power
grid.
Fig. 22.2: ECG sensor (ADS1292R) block diagram.
22.3 Experimental setup
We developed the experimental protocol which operation is described below. Four

healthy volunteers were recruited in this study. We organized the experiments in the
following steps:
1. Subjects were asked to wear the developed prototype. Four ECG electrodes
were placed symmetrically on the body. Central Acquisition Unit was fixed
below the chest. The volunteer with wearable sensors is shown in the 22.3
2. Measurements were taken in AGH Wearables Research Group Laboratory. Each
subject was asked to perform for about 20s the following activities: sitting,
walking and squatting. Measurements were repeated to obtain 17 data files -
each activity saved in separated file.
3. ECG data were recorded using manufacture’s software, while the rest of the
data was recorded on the SD card using developed device.
Fig. 22.3: The volunteer with wearable sensors.
22.4 Methods
22.4.1 4.1 Dataset preparation and processing
During preprocessing, we focused on signal filtering and calculating features in de-

fined time window. We randomly split acceleration signal database files with pro-
portion 0.8:0.1:0.1 into train set, validation set and test set. Then, we applied a 3
seconds moving window for each signal for features extraction. This aimed to ob-
tain between-individual generalization and larger dataset, which would be sufficient
for machine learning models. Finally we obtained the following classes: idle, walk,
squaatting:
• training set: 54 108 34 segments
• validation set: 51 55 18 segments
• testing set: 19 53 20 segments
Validation set was used to check, if the model is prone to overfitting at each training
iteration. In turn, the testing set was utilized to calculate final result, which is pre-
sented in this paper. As representative features, we decided to calculate mean, stan-
dard deviation, peak-to-peak amplitude and number of mean crossing. The heart rate
based on ECG measurement certainly provides vital information about patient’s ac-
tivity. However, during experiments it occurred, that this parameter features signifi-
cant inertia during performing exercises. In other words, the patient’s pulse changes
with a latency being too considerable in terms of quick movement classification. As
heart rate introduced data redundancy, we used only acceleration signals for further
analysis and classification.
22.4.2 4.2 Classification with machine learning algorithms.
For the purpose of this paper, we decided to evaluate following machine learning
algorithms: Softmax Regression 22.4, Multilayer Perceptron 22.5, SVM and Ran-
dom Forest as their proven efficacy was widely stated. We were training our models
on training set, and checking them for overfitting on validation set in each iteration.
Overall result was calculated on testing set.
Fig. 22.4: Softmax Regression model diagram.
As far as MLP is concerned, we tested 3 and 6 neurons in hidden layer. The result
shows that 3 neurons are not enough to memorize all possibilities given in training
time. The 6 neurons occurred to be the better solution, however with this model we
met an issue of overfitting, which had to be solved by adding a L2 regularization.
When it comes to random forest algortihm, I has appeard that set of two trees with
depth of 5 have been enough to represent chosen classes properly.
Fig. 22.5: Representation of Multilayer Perceptron with 6 neurons in hidden layer.

Detection rate was improved, due the a nonlinearity provided by hidden layer.
Fig. 22.6: Loss function in epoch for MLP with 6 neurons in hidden layer.
Fig. 22.7: Loss function in

epoch for MLP with 6 neurons in hidden layer. The L2 regularization was performed
to avoid overfitting.
We noticed that chosen features are not linearly separable. The reason we are able
to draw this conclusion is that MPL and Random Forest non-linear classifiers had
the best scores. SVM and Softmax Regression are both linear classifiers and their
outcomes confirm our suspicions. The L2 regularization significantly increased the
training efficacy which is depicted in the 22.6. and 22.7. What is more, great result
of random forest classifier and its low computional complexity for searching spec-
ified element (log(n), where n is number of leafs) makes it even better solution for
embedded devices than mlp. We can conclude, that random forest will be our first
try for implemented classifier on emebedded device.
To effectively evaluate proposed algorithms in terms of activity classification, we

calculated sensitivity (Se) and positive predictivity (pp) for each classifier. We also
attached f1 (22.1) score which combine both Se and pp:
f 1 = 2 ∗ Se ∗ pp/(Se + pp) (22.1)

Sensitivity means how many occurence of single class were detected in compar-
ison with all occurrence of this class in signal. It is expressed by the formula (22.2):
Se = T P/(T P + FN) (22.2)

Positive predictivity, on the other hand, is expressed as (22.3):
pp = T P/(T P + FP) (22.3)

where:
TP – true positive
FP – false positive
FN – false negative
It describes how many episodes have been detected correctly in comparison with
all episodes which were denoted as true by classifier. The calculated parameters for
each classifier are listed in the table below 22.1:
Table 22.1: The set of parameters calculated for each classification algorithm.
Classifier Se pp f1
SR 0.92 0.91 0.91
MLP (3 neurons) 0.92 0.92 0.92
MLP(6 neurons, L2 regularization) 0.95 0.93 0.94
SVM 0.8 0.76 0.74
RF 0.99 0.99 0.99
Based on the developed feature vectors, we performed numerous classification

experiments using different machine learning models. We found that rehabilitation
activity can be detected with 99% with the use of Random Forest algorithm. The
second best result was obtained for MPL with 6 neurons in hidden layer, however in
this case a regularization was needed to avoid overfitting of the neural network clas-
sifier. SMV and Softmax regression methods perform worse effectiveness. This may
be motivated by the linear characteristics of these algorithms, whilst the extracted
features can only be separated by non-linear models. Certainly an optimization of
the future vector in conjunction with bigger data set may provide even better results
for described application.
Above results prove the legitimacy of using proposed methodology for classifi-
cation of rehabilitation exercises. The designed wearable prototype may be success-
fully used in monitoring biomedical signals. The intuitive reconfiguration of the
system in conjunction with possible adaptation of another sensors make it suitable
for evaluation purposes. The 99% of correctly recognized activities in case of RF

gives motivation for further research. Moreover, this method features the BigO com-
putational complexity of log(n), which is the best score within tested algorithms.
The fusion of the best classification rate with high computational efficacy makes
the Random Tree a reasonable solution for embedded implementation on wearable
devices. This approach would overcome a necessity of processing the data on per-
sonal computer or remote server, providing the on-chip implementation of the whole
assisted living system.
22.6 Acknowledgment
This project was funded by the National Center for Research and Development
based on the decision number TANGO1/270395/NCBR/2015.
References
1. Chan, M., EstÃĺve, D., Fourniols, J.-Y., Escriba, C., Campo, E.: Smart wearable systems:
Current status and future challenges. Artif Intell Med. 56, 137–156 (2012).
2. Klingeberg, T., Schilling, M.: Mobile wearable device for long term monitoring of vital signs.
Comput. Methods Programs Biomed. 106, 89–96 (2012).
3. Kantoch, E., Ieee: Technical Verification of applying Wearable Physiological Sensors in
Ubiquitous Health Monitoring. In: 2013 Computing in Cardiology Conference. pp. 269–272
(2013).
4. Kańtoch, E.: Telemedical human activity monitoring system based on wearable sensors net-
work. In: 41st Computing in Cardiology Conference, CinC 2014. pp. 469–472 (2014).
5. Huang, B., Giggins, O., Kechadi, T., Caulfield, B.: The limb movement analysis of rehabilita-
tion exercises using wearable inertial sensors. In: 2016 38th Annual International Conference
of the IEEE Engineering in Medicine and Biology Society (EMBC). pp. 4686–4689. IEEE
(2016).
6. Lu, J., Zhang, T., Sun, Q., Kadiwal, S., Unwala, I., Hu, F.: Monitoring of Paces and Gaits
Using Binary PIR Sensors with Rehabilitation Treadmill. Med. . . . . 1, 5315–5318 (2016).
7. Saracino, L., Ruffaldi, E., Graziano, A., Avizzano, C.A.: Fusion of wearable sensors and mo-
bile haptic robot for the assessment in upper limb rehabilitation, (2016).
8. Pierleoni, P., Belli, A., Maurizi, L., Palma, L., Pernini, L., Paniccia, M., Valenti, S.: A Wear-
able Fall Detector for Elderly People Based on AHRS and Barometric Sensor, (2016).
9. Gravina, R., Alinia, P., Ghasemzadeh, H., Fortino, G.: Multi-sensor fusion in body sensor
networks: State-of-the-art and research challenges. Inf. Fusion. 35, 68–80 (2017).
10. NEITZEL, Jennifer A.; DAVIES, George J. The Benefits and Controversy of the Parallel Squat
in Strength Training and Rehabilitation. Strength & Conditioning Journal, 2000, 22.3: 30.
Part III
Modelling and simulations in biomechanics
Chapter 23
Assessment of balance of older people living at
a social welfare home
Katarzyna Jochymczyk-Woźniak1 , Katarzyna Nowakowska1 , Robert Michnik1 ,

Agnieszka Nawrat-Szołtysik2 and Wioletta Górka3
Abstract. More frequently fall of older people can be caused by increasing of the
degeneration of a human motor system and a human maintaining balance system.
The aim of study was to determine the association of balance and risk of falling of
older people lived at a social welfare home. The authors examined the dependence
between the ability to maintain balance in the upright position and the assessment
of balance and risk of falling in a functional way. The study group were consisted
of 26 elderly people. Postural stability examinations was carried out based on the
Romberg test, while the functionally balance was evaluated with a Up&Go test. The
statical tests did not manifest significant correlation between the static and dynamic
stabiliography parameters, but one proved that there existed strong differences for
obtained ellipse area results, among the group of patients performing the Up&Go
Test within shorter and longer time than 20 sec.
Keywords: stabilography, path length, elipse area, Zebris-FDM, older people
23.1 Introduction
Equilibrium is a postural system that preserves vertical orientation due to bal-

ancing of opposing forces acting on a body and their torques. The equilibrium is
ensured by the nervous system with reflexive tension of adequate muscles known
as postural and antigravitational ones. Keeping the balance, both static and dynamic
Department of Biomechatronics, Faculty of Biomedical Engeenering, Silesian Univesity of

Technology, Poland,
e-mail: Katarzyna.Jochymczyk-Wozniak@polsl.pl,
WWW home page: https://www.ib.polsl.pl · Chair of Nervous and Locomotor Sys-
tem Physiotherapy, Academy of Physical Education in Katowice, Poland · Students Scientific
”BIOKREATYWNI”, Faculty of Biomedical Engeenering, Silesian Univesity of Technology,
Poland

https://doi.org/10.1007/978-3-319-70063-2_23
218 Katarzyna Jochymczyk-Woźniak et al.
[7] (especially during the movement), depends on the proper control system func-
tioning of the posture holding the center of gravity projection of the body within the
underpin field and being able to counteract the external forces that may destabilize
the posture [9], [10], [11].
The elderly’s systems responsible for keeping the body equilibrium are getting
worse due to the lowered nervous system efficiency and its gradual degradation. The
functioning of senses is getting worse as well. The lowering of peripheral excitabil-
ity of the part of atrial organ and sense touch body disorder occur. Sharpness and
both peripheral and spatial ability vision are getting impaired. A significant decrease
in muscle mass and efficiency is observed. Increased muscles rigidity and trembling
growth dominate in the clinical picture, which is associated with the circumstances
change of the control balance posture [2]. Neuromotor activities get moderated, the
time response lengthens, one worsens: perceptiveness, divisibility and motor coor-
dination [14]. The most often consequence of the elderly’s balance loss are falls. As
the surveys show about 30-40% of healthy, independent individuals over 60 fall at
least once a year, whereas almost 50% over 80. In Poland for every 100 individuals
there are 47 downfalls among people aged 70-74 and this number increases accord-
ingly to age [3], [5], [16]. Even 10-25% of downfalls lead to fractures affecting the
patients of the social assistance houses and hospitals. These falls are three times
more often comparing to the patients dwelling individually. The residence change
of social assistance house patients (e.g. moving to different room) increases this
hazard by 50% [8]. The majority of the elderly’s downfalls occur during moving.
Among all downfalls 60% of them are observed towards forward movement because
of the slip [2], [14]. One of the most promising diagnostics’ method is posturogra-
phy that allows to assess the magnitude of human balance boundries’ changes and
plethora of pathological states handicapping the control of upright position balance
[3], [16]. As far as diagnostics is concerned there are important tests assessing the
balance in functioning way (e.g. Up&Go Test, Berg’s equilibrium scale, Tinetti’s
test etc.), which allow to observe moving demeanor in postural challenging circum-
stance, simultaneously stating the base downfall hazard based on the tests standards
[4].
Taking into consideration above mentioned the aim of the research is the assess-
ment of the balance and the downfall hazard of the elderly dwelling in social as-
sistance houses, based on stabilography parameters (track length, ellipse area) and
Up&Go Test. Moreover, the researches decided to state the dependence between the
ability of keeping the balance in upright position and both balance assessment and
the downfall hazard in functioning way.
23.2 Methods
The experimental research was being carried out in Saint Elizabeth Welfcare
Social Centre located in Ruda Ślaska
˛ using the Zebris FDM system dedicated to
evaluate the upright balance position. The investigation of the body balance posture
23 Assessment of balance 219
(the analysis of resultant force’s application point of the components acting between
feet and ground) was based on the Romberg’s test assessing the balance during the
free standing on both lower limbs with eyes respectively: closed and open. Legs
astride within the pelvis width and arms arranged freely along the body. There were
following quantities analysed: path length (the distance that GOP covers) and the
ellipse area (COP displacement area). The balance was being assessed in functional
way with Up&Go Test. The test consists in performing some simple tasks such as:
rising form the chair holding upright back, covering the flat distance of 3 meters,
crossing the line ending a given sector, performing 180◦ rotation, returning to the
chair and installing the sitting position. The seat height (46 cm) was established
according to the research findings of Siggeirdottir [13]. The test performing time
was measured with the command ’Start’ until installing the sitting position again
[12], [13]. The investigation was carried out twice with a measure of average value.
The research group consisted of the elderly aged 65–94 (av. value: 79 ± 8, body
mass: 69±1 kg, body height: 159±6 cm) respectively 17 females (age: 82±6, body
mass: 68 ± 16 kg, body height: 158 ± 5 cm) and 9 males (age: 73 ± 6, body mass:
74±8 kg, body height: 161±6 cm). The main acceptable condition was the age over
65 and the shortage of neurological occurrences. Men and women who understood
the essence of the test were qualified. According to established norms for Up&Go
Test people over 65, characterized by good functional mobility, were informed to
perform the test within 10-20 sec [14], [15]. The persons being tested, who perform
the test over 20 sec, are vulnerable to mobility impairment and downfalls. Hence,
the research group was splitted into two subgroups: patients performing the test up
to 20 sec (group quantity: 15, av. age: 77 ± 8, body mass: 67 ± 12 kg, body height:
158±6 cm), and over 20 sec (group quantity: 11, av. age: 80±6, body mass: 73±16
kg, body height: 160 ± 6 cm).
The obtained results of Up&Go Test and Romberg’s Test were presented in
Table 1.
Table 23.1: Descriptive statistical analysis of obtained results
Up&Go Test [s] Path length [mm] Elipse area [mm2 ]

mean 20,37 499,07 116,40
std deviation 6,95 150,15 104,42
median 18,50 434 78
max 38,20 984 464
min 12,00 324 13
Fig. 23.1: Comparison of the received path length values with Jurgens et al. results
[6]
Fig. 23.2: Comparison of the received elipse area values with Baltich et al. and
Merlo et al. results [1], [8]
The obtained values of parameters of static stabiliography for the elderly’s given
group were compared with the findings of other authors [1], [6], [8]. The obtained
values of the path length, that is covered by the center of gravity within 30 sec,
correspond to the findings of Jorgensen et al. [6]. However, the range of ellipse area
values obtained in own research is much lower comparing to research of the elderly
carried out by Marlo et al. [8] and Baltich et al. [1] (Fig. 1-2).
The individuals qualified to the research group were performing the Up&Go Test
with average time of 20.37 sec. In the research group consisting of 26 eldery people,
42% of the patients were performing the test within the time longer than 20 sec. One
can suppose that less than half of the group is vulnerable to often downfalls. Thus,
the results were divided, according to the time of performing Up&Go Test, where
the limit time value was 20 sec. In order to state the normal distribution of obtained
findings, the Shapiro-Wilk test carried out. It was shown that analysed parameters
did not comply with normal distribution. The distribution of obtained parameters
values of static stabiliography in the function of time performing Up&Go Test was
presented in Figure 3 and Figure 4.
In order to evaluate the dependency between holding balance in upright position
and balance assessment is functional way, one determined the Spearman correlation
coefficient, i.e. the number determining how given variables are correlated. In ac-
cepted time intervals (up to 20 sec and over 20 sec), one calculated the correlation
coefficient between time Up&Go Test in function of path length and ellipse area.
The anlyses that had been carried out did not statically evidence any significant cor-
relation between static and dynamic stabiliography parameters. The (statical) influ-
ence of time performing Up&Go Test (up to 20 sec and above 20 sec) on considered
static stabiliography parameters (path length and ellipse area) was investigated with
nonparametric test U Mann-Whitney for independent samples. Having assumed the
significance level α = 0, 05 one demonstrated that there were differences, statically
important, according to obtained results of the ellipse area between the analysed
groups (p = 0,038).
Fig. 23.3: The distribution of obtained results of path length in the function of time
performing Up&Go Test
Fig. 23.4: The distribution of obtained results of elipse area in the function of time
performing Up&Go Test
23.4 Conclusion
In the article the dependencies between the ability to hold the balance in upright
position and hazard assessment of downfall for the elderly were analysed. The static
stabiliography research was being carried out with Zebris FDM platform, whereas
the dynamic balance assessment was based on Up&Go Test [7]. Interpreting the
Up&Go Test results one assumed that fit persons being tested performed the test
within 20 sec. That amount of time differentiates individuals with good functional
mobility and those who are vulnerable to downfalls. The persons being tested who
performed the test in 20 sec or longer can be classified as those revealing balance
disturbances facing problems with altering the sitting onto upright position and vice
versa. The average time of performing the test in a given group was 20 sec. There
were 42% of the persons being tested to be found in the group that was vulnera-
ble to downfalls. The average value of the path length of the persons being tested
is 499,07 mm and the ellipse area is 116, 39mm2 . The statical tests, that had been
carried our, did not manifest significant correlation between the static and dynamic
stabiliography parameters, but one proved that there existed strong differences for
obtained ellipse area results, among the group of patients performing the Up&Go
Test within shorter and longer time than 20 sec (p = 0,038). The entire research
follows, therefore, that there is not strong dependency between holding the balance
in upright position and functional assessment. It seems to be necessary to imple-
ment both investigations simultaneously (static stabiliography, Up&Go Test). The
research, being discussed in this article, indicates that early diagnosis of balance
impairment is extremely important, because combined with well oriented rehabili-
tation can mitigate the negative aspects of malfunction and improve the mobility of
the elderly. Thus determining the hidden causes of balance loss and associated with
it the downfall hazard is an essential challenge.
References
1. Baltich J., von Tscharner V., Zandiyeh P. and Nigg B.M.: Quantification and reliability of center
of pressure movement during balance tasks of varying difficulty. Gait&Posture, 40, 327–332
(2014)
2. Błaszczyk J.W.: The method of spatial distribution histograms and contour plots applied to
postural stability evaluation in young and elderly subjects, in: From Basic Motor Control to
Function Recovery (eds. Gantchev N. and Gantchev G. N.), Academic Publishing House, Sofia,
197-202 (1999)
3. Błaszczyk J., Czerwosz L.: Postural stability in the process of aging. Gerontologia Polska,
13(1), 25–36 (2005) (in Polish)
4. Bohannon R.: Reference values for the timed up and go test: a descriptive meta-analysis. Journal
of Geriatric Physical Therapy, 29(2), 64–68 (2006)
5. Czerwiński E., Białoszewski D., Borowy P., Kumorek A. and et al.: Epidemiology, Clinical Sig-
nificance, Costs and Fall Prevention in Elderly People. Ortopedia Traumatologia Rehabilitacja,
10(5), 419–428 (2008) (in Polish).
6. Jorgensen M.G., Rathleff M.S., Laessoe U., Caserotti P., Nielsen O.B.F., Aagaard P.: Time-of-
day influences postural balance in older adults. Gait&Posture, 35, 653–657 (2012)
7. Kostiukow A., Rostkowska E., Samborski W.: Assessment of postural balance function. An-
nales academiae medicae stetinensis, 55, 102–109 (2009)
8. Merlo A., Zemp D., Zanda E., Rocchi S., Meroni F., Tettamanti M., Recchia A., Lucca U.,
Quadri P.: Postural stability and history of falls in cognitively able older adults: The Canton
Ticino study. Gait&Posture, 36, 662–666 (2012)
9. Michnik R., Jurkojć J. Wodarski P., Gzik M., Jochymczyk-Woźniak K., Bieniek A.: The influ-
ence of frequency of visual disorders on stabilographic parameters. Acta of Bioengineering and
Biomechanics, 18(1), 25–33 (2016)
10. Michnik R., Jurkojć J., Wodarski P., Gzik M., Bieniek A.: The influence of the scenery and the
amplitude of visual disturbances in the virtual reality on the maintaining the balance. Archives
of Budo, 10, 133–140 (2014)
11. Ocetkiewicz T., Skalska A., Grodzicki T.: BBalance estimation by using the computer balance
platform: repeatability of the measurements. Gerontologia Polska, 14(1), 144–148 (2006) (in
Polish)
12. Shumway-Cook A., Brauer S., Woollacott M.: Prediciting the probability for falls in
community-dewelling older adults using the Timed Up&Go Test. Physical Therapy Journal,
80(9), 896–903 (2000)
13. Siggeirdottir K., Jonsson B.Y., Jonsson H., and et al.: The Timed "Up&Go" is depended on
chair type. Clinical Rehabilitation, 16(6), 609–616 (2002)
14. Szot P., Golec J., Szczygieł E. and et al.: Overview of selected functional tests used in as-
sessment of the risk of falls of older persons. Gerontologia Polska, 16(1), 12–17 (2008) (in
Polish)
15. Szpringer M., Wybraniec-Lewicka B., Czerwiak G. and et al.: Falls and injuries in geriatric
age. Studia Medyczne, 9, 2008, 77–81 (2008) (in Polish)
16. Wiszomirska I., Kaczmarczyk K. and Ilnicka L.: The effect of training stimulation vestibular
organ on postural stability in elderly. Post Rehab, 4(5/10), 5–10 (2010) (in Polish)
Chapter 24
Assessment of locomotor functions of patients
suffering from cerebral palsy qualified to treat
by different methods
Katarzyna Jochymczyk-Woźniak1 , Katarzyna Nowakowska1 , Robert Michnik1 ,

Agnieszka Konopelska2 , Jerzy Luszawski3 and Marek Mandera3
Abstract. The aim of this study was to assess locomotor functions of patients with
cerebral palsy and qualified for various treatment procedures on the basis of indica-
tive methods and parameters constituting the Gillette Gait Index. The studied group
consisted of 57 healthy children and 60 children with different types CP and dif-
ferent degrees of disability: REH - patients subjected to physiotherapy only, BOT
- patients subjected to the injection of botulinum toxin and intensive rehabilitation
and RIZ - patients qualified for the surgical procedure of the selective dorsal rhizo-
tomy. Gait investigations were carried out using BTS Smart system and the gait as-
sessment was made on the basis of gait indices methods (GGI, GDI). The statistical
analysis demonstrated that the most diversified parameters characterising patients
with CP qualified for various treatments are: knee flexion at initial contact, time of
peak knee flexion, range of knee flexion, range of pelvic tilt.
Keywords: botulinum toxin (BOT), gait analysis, gait indicase methods GGI, GDI,
physiotherapy (REH), the selective dorsal rhizotomy (SDR)
24.1 Introduction
Cerebral palsy (CP) is concerned with brain damage (particularly in relation

to the central motor neuron) which may occur during pregnancy, in the perinatal
or postnatal period, while the brain continues to develop. Cerebral palsy is charac-
terised by various disorders affecting the motor system and impaired posture control.
In the general children’s population, cerebral palsy affects two in 1000 children born
Department of Biomechatronics, Silesian University of Technology, Roosevelta 40 41-800 Zabrze,

Poland
e-mail: Katarzyna.Jochymczyk-Wozniak@polsl.pl,
WWW home page: http://www.ib.polsl.pl · Institute of Theoretical and Applied Me-
chanics, the Silesian University of Technology, Konarskiego 18a, 44-100 Gliwice, Poland · Upper
Silesian Children’s Health Centre in Katowice, Medyków 16, 40-751 Katowice, Poland

https://doi.org/10.1007/978-3-319-70063-2_24
worldwide. Fifty percent of patients diagnosed with CP suffer from an excessive in-
creased muscle tone referred to as spasticity and defined as excessive movement
velocity-dependent reflex resistance to the passive stretch of muscles. The genera-
tion of contractures restricting movement in joints and responsible for the secondary
deformation of joints pose a serious problem for patients with long-lasting limb
spasticity. Traditional methods used when treating cerebral palsy include physio-
therapy, an ancillary treatment such as pharmacotherapy (e.g. botulinum toxin in-
jections or the implantation of a baclofen pump), the use of orthopaedic aids and the
application of alternative therapeutic methods (e.g. the use of space suits or hyper-
baric oxygen therapy). Modern neurorehabilitation takes advantage of various de-
vices aiding in the re-education of gait, e.g. exoskeletons. Some cases require opera-
tive treatments involving orthopaedic procedures or, in exceptional cases, neurosur-
gical procedures [1], [2]. The use of botulinum toxin in fighting spasticity consists
in the injection of reversible nerve-muscle blockade aimed to weaken muscles. One
of operative methods for treating the spasticity of lower limbs is the selective dor-
sal rhizotomy (SDR). This neurosurgical operation involves cutting several selected
fibres of the sensory nerves between the muscles and spinal cord (sensory nerves
transmit sensory information from muscle spindles and other parts of the body). The
reduction of the spasticity of lower limbs performed using the above-named method
improves locomotor functions and the posture [10]. The assessment of the motor
organ, and in particular the assessment of locomotor functions of patients suffering
from cerebral palsy are based on subjective (i.e. clinical) and (increasingly often)
objective (i.e. biomechanical) tests. The biomechanical assessment is primarily fo-
cused on the analysis of gait as this form of locomotion constitutes an important
element enabling the satisfaction of daily life needs. Presently, the assessment of
gait model disorders utilises systems for the triplanar analysis of gait. Previously
conducted gait tests enable the analyses of time-space parameters, kinematic quan-
tities (e.g. courses of angles in individual joints during a single gait cycle), reactions
of the base, resultant momenta of muscular forces in joints of lower limbs as well
as muscle activity [3], [4], [6]. In addition, the objective analysis of gait enables
the determination of the so-called GGI (The Gillette Gait Index) and GDI (The Gait
Deviation Index) gait indices allowing the interpretation of results using a single nu-
merical value. This value reflects the gait subjected to examination and demonstrates
how it differs from the mean of regular gait [3], [9], [10], [11]. The qualification of
cerebral palsy patients for appropriate therapeutic procedures faces numerous dif-
ficulties, yet the objective triplanar analysis of gait is a very useful diagnostic tool
facilitating the decision-making process. Neither Polish nor foreign available ref-
erence publications contain information concerning the statistical analysis of gait
test results related to patients qualified for operative selective dorsal rhizotomy or
other therapeutic methods. Due to the foregoing, this study aims to assess locomotor
functions of patients diagnosed with cerebral palsy and qualified for various treat-
ment procedures on the basis of indicative methods and parameters constituting the
Gillette Gait Index.
24 Assessment of locomotor functions of patients suffering from cerebral palsy 227
This paper is concerned with the research assessing the locomotor functions of
patients with CP treated by means of various methods. The studied group was com-
posed of 57 healthy children at the age of 7 - 17 and 60 children at the age of 3 -
17 with different types CP and different degrees of disability. Some patients were
able to move independently, some used ancillary equipment (such as crutches or
a Zimmer frame). Due to different treatment methods, the patients were divided
into three groups: the first group: REH - patients subjected to physiotherapy only
(10 participants), the second group: BOT - patients subject to the injection of bo-
tulinum toxin and intensive rehabilitation (35 participants) and the third group: RIZ
- patients qualified for the surgical procedure of the selective dorsal rhizotomy (15
participants). Five patients qualified for treatment with selective dorsal rhizotomy
use crutches every day but while gait tests all patients were moving independently
(without orthopedic equipment). It was one of the conditions for joining the patient
to study group. The testing methodology was divided into several stages. The stages
are presented in a block diagram in Figure 1.
Fig. 24.1: Simplified diagram of the process applied to diagnose locomotor func-
tions of patients with CP
Measurements were conducted in the Upper Silesian Children’s Health Centre in

Katowice. The assessment of locomotor functions was done on the basis of clinical
and biomechanical tests. Gait investigations were carried out using a system for
the triplanar analysis of movement - BTS Smart system. The system consisted of
8 optoelectronic cameras which enabled the recording of the movement (recording
of the location of passive markers) in three planes as well as two dynamometric
platforms of the Kistler company. The markers were placed on the patients’ bodies
according to Davis’ protocol. For each patient there were recorded about 20 dynamic
tests whereas the analysis has been subjected to about ten cycles of gait for each
patient.
The gait assessment was made on the basis of gait indices methods (GGI, GDI)
using the author’s applications written in Matlab environment. The Gillette Gait
Index (GGI) is based on 16 kinematic parameters (time of toe off (P1), walking
speed (P2), cadence (P3) mean pelvic tilt (P4), range of pelvic tilt (P5), mean pelvic
rotation (P6), minimum hip flexion (P7), range of hip flexion (P8), peak abduc-
tion in swing (P9), mean hip rotation in stance (P10), knee flexion at initial contact
(P11), time of peak knee flexion (P12), range of knee flexion (P13), peak dorsiflex-
ion in stance (P14), peak dorsiflexion in swing (P15), mean foot progression angle
(P16) which according to clinicians accurately describe the gait of patients with
CP. This index is expressed by means of a single numerical value. The value of the
GGI index is a measure of the difference between a set of discrete parameters of
the gait of a given patient and mean parameters of the gait of healthy individuals
[9]. Jochymczyk-Woźniak [3] found in her work that a normative value of the GGI
index equalled 15.71 (7.46 – 30.00). The GDI takes into consideration 9 variable
kinematic values describing a range of angles in the knee, hip, pelvic and tarsal
joints [10]. The data are collected every 2% of the gait cycle, which amounts to 459
measuring points for one patient. The determined GDI value provides information
on the difference of the gait of a given patient from regular gait. The gait of a tested
person is similar to the gait of the control group, where the GDI ≥ 100, whereas
each change in the GDI index by 10 below 100, means one standard deviation from
the mean determined for the control group. For each of the sixteen selected param-
eters making up the Gillette Gait Index the ranges of classification were determined
starting from very good, through acceptable, to weak and very weak. Figure 2 shows
the classification of the results according to the scale adopted. The GGI and GDI re-
sults obtained from the studied group of patients were compared by means of the
single-factor Analysis of Variance - ANOVA as well as the Kruskal-Wallis test.
Fig. 24.2: Classification of results acc. to the scale adopted

24.3 Results
The experimental tests were conducted on patients with CP who were qualified
for the following types of treatment: rehabilitation (REH), botulinum toxin (BOT),
and surgical procedure of selective dorsal rhizotomy (RIZ). On the basis of these
tests, space-time parameters as well as kinematic and dynamic parameters were de-
termined. The angles of particular joints of lower limbs and pelvis are the most
often analysed kinematic values during the gait. Figure 3 show mean values, stan-
dard deviation, median, maximum and minimum values of the GGI and GDI indices
calculated for the group of patients qualified for treatment with the use of botulinum
toxin and selective dorsal rhizotomy as well as patients rehabilitated in a process of
individual therapy.
The conducted statistical analyses proved that only in the case of four parameters
(mean pelvic rotation (P6), range of hip flexion (P8), mean hip rotation in stance
(P10), peak dorsiflxion in stance (P14), the differences of the gait indices between
the groups were not statistically significant (p≤0.05).
Fig. 24.3: Results of GGI and GDI indices for the group of patients with cerebral
palsy
Figure 4 presents mean values of 16 selected parameters which were determined

for individual groups of patients according to the scale adopted.
24.4 Discussion
Within the framework of the research the Gilette Gait Index and Gait Deviation
Index values were determined for patients with cerebral palsy who were qualified
for treatment with different methods. The mean GGI value for patients qualified for
the surgical procedure of selective dorsal rhizotomy equalled 1118.53, and the mean
GDI value was 61.19. Big dispersion of the obtained values results from the fact that
Fig. 24.4: Mean values of 16 parameters determined for three groups of patients
according to the scale adopted
the studied group includes patients having different degrees of disability and thus
different degrees of muscular tension. Equally big diversity of the results obtained
for individual indices may be observed in the case of patients who were qualified
for the injection of botulinum toxin into selected groups of muscles. The mean GGI
value amounted to 350.71, while the GDI equalled 73.89. In the group of patients
who were rehabilitated only by means of motor rehabilitation a considerably lower
mean value of the Gillette Gait Index was obtained, namely 115.53, whereas the
mean GDI value amounted to 88.19. GGI and GDI values obtained in own reaserch
for patients who were qualified for treatment using rehabilitation and botulinum are
similar with the results obtained by other researchers [5], [8], [9].
The study-related work involved the statistical analysis (ANOVA, the Kruskal-
Wallis test) of the individual parameters constituting the GGI in patients qualified
for various treatments. On the basis of the value of probability level p, it was pos-
sible to notice that major parameters diversifying individual groups included knee
flexion at initial contact (P11), time of peak knee flexion (P12), range of knee flex-
ion (P13), range of pelvic tilt (P5). The aforesaid observation is demonstrated in the
diagrams presented in Figure 5. It was observed that all the parameters concerning
the knee joint were more diversified in individual groups of patients. This led to the
conclusion that the analysis of knee joint kinematics provides valuable information
when qualifying patients for appropriate treatment methods, which was also con-
firmed by Roberts at al. [7]. The analysis of parameters characterising gait revealed
that in approximately 90% of patients qualified for the operative SDR, the values
of parameters characterising the knee joint differed significantly from the normative
values. One of the primary problems of patients qualified for the SDR is spasticity
limiting the mobility of individual joints, which as regards the knee, is manifested
by the joint rigidity when walking and can be ascribed to the lack of strength in
Fig. 24.5: The values of parameter most differentiating different groups of patients
the shank in the pre-swing phase as well as to the spasticity of thigh muscles, pri-
marily of the straight muscle of the thigh. The lack of flexion in the knee joint in
the pre-swing phase is the key parameter when qualifying patients for the SDR [7].
The tests conducted enabled the assessment of locomotor functions on the basis of
gait indices methods. The GGI and GDI were used for determining deviations of CP
patients’ gait from that of healthy individuals. The single-factor analysis of variance
(ANOVA) revealed that the differences in the GDI results were statistically relevant
for 3 groups of patients qualified for various treatments (the comparison of 3 in-
dependent groups) (p<0.05). The performed Kruskal-Wallis test (p<0.05) revealed
that the differences between the GGI results for three groups of CP patients were
also statistically relevant. Statistically relevant differences were also revealed as re-
gards values of indices characterising locomotor functions of CP patients qualified
for various treatments. The methodology developed can be used for the objective
and precise qualification of patients for individual methods of treatment.
24.5 Conclusion
The tests conducted enabled the assessment of locomotor functions on the ba-
sis of gait indices methods. These methods offer a very precise and accessible tool
for comparing the results of CP patients qualified for various treatments. The tests
demonstrated that the greatest deviations of the GGI and GDI (from the normative
data) characterised the third group. In turn, in the first group, these indices were less
different from the values characterising healthy children. The statistical analysis of
16 parameters constituting the GGI revealed that the most diversified parameters
characterising patients qualified for various treatments included knee flexion at ini-
tial contact (P11), time of peak knee flexion (P12), range of knee flexion (P13),
range of pelvic tilt (P5). The foregoing justifies a conclusion that the analysis of
knee joint kinematics should be taken into consideration when qualifying patients
for specific treatment methods.
Acknowledgements
The study was supported by research grant no. DEC-2011/01/B/NZ7/02695 of the

Ministry of Science and Higher Education in Poland
References
1. Brochard S., Remy-Neris O., Filipetti P., Bussel B.: Intrathecal baclofen infusion for ambulant
children with cerebral palsy. Pediatr Neurol, 40(4), 265–70 (2009)
2. Desloovere K., De Cat J., Molenaers G., Franki I., Himpens E., Van Waelvelde H., Fagard
K., Van den Broeck C.: The effect of different physiotherapy interventions in post-BTX-A
treatment of children with cerebral palsy. Eur J Paediatr Neurol, 16(1), 20–28 (2012)
3. Jochymczyk-Woźniak K., Janoska P., Michnik R.: Evaluation of selected parameters in the gait
of children after removing the posterior cranial tumor (Ocena wybranych parametrów chodu
dzieci po usuni˛eciu guza tylnej jamy czaszki), in: Selected issues of sport biomechanics and
rehabilitation (Wybrane zagadnienia biomechaniki sportu i rehabilitacji), T. 1. Red. Czesław
Urbanik, Andrzej Mastalerz. Warszawa : Akademia Wychowania Fizycznego Józefa Piłsud-
skiego, 57–68 (2012) (in Polish)
4. Michnik R., Jurkojć J., Pauk J.: Identification of muscles forces during gait of children with
foot disabilities. MECHANIKA, 6(80), 48–51 (2009)
5. Molloy A., McDowell B.C., Kerr C., Cosgrove A.P.: Further evidence of validity of the Gait
Deviation Index. Gait & Posture, 31, 479–482 (2010)
6. Nowakowska K., Michnik R., Jochymczyk-Woźniak K., Jurkojć J., Mandera M., Kopyta I.:
Application of gait index assessment to monitor the treatment progress in patients with cerebral
palsy (in:) Information Technologies in Medicine 5th International Conference, ITIB, Kamień
Ślaski,
˛ Poland, June 20-22, 2016 Proceedings, Pi˛etka E., Badura P., Kawa J., Wieclawek W.
(eds.), Advances in Intelligent System and Computing 472, vol. 2, 75–85 (2016)
7. Roberts A., Stewart C., Freeman R.: Gait analysis to guide a selective dorsal rhizotomy pro-
gram. Gait & Posture, 42, 16–22 (2015)
8. Romei R., Galli M., Motta F., Schwartz M., Crivellini M.: Use of the normalcy index for the
evaluation of gait pathology. Gait & Posture, 19(1), 85–90 (2004)
9. Schutte L.M., Narayanan U., Stout J.L., Selber P., Gage J.R., Schwartz M.H.: An index for
quantifying deviations from normal gait. Gait & Posture, 11, 25–31 (2000)
10. Schwartz M., Rozumalski A.: The gait deviation index: A new comprehensive index of gait
pathology. Gait & Posture, 28, 51–357 (2008)
11. Syczewska M, Dembowska-Bagińska B, Perek-Polnik M, Kalinowska M, Perek D.: Gait

pathology assessed with Gillette Gait Index in patients after CNS tumour treatment. Gait &
Posture, 32, 358–362 (2010)
Chapter 25
Body part accelerations evaluation for chosen
techniques in martial arts
Sebastian Glowiński, Andrzej Blażejewski, Tomasz Krzyżyński
Abstract. The authors have applied microelectromechanical systems (MEMS) for

the motion analysis in sports activities by attaching them onto the judoist body seg-
ments. The inertia sensor has been attached onto the most often injured body re-
gion as head, torso and legs. MEMS can measure acceleration and provide real-time
feedback to judoists so they can perfect their form. Therefore, authors proposed the
judoist’s skill evaluation and techniques harmfulness, extracting the body segment
accelerations. The purpose of this study was to assess the acceleration profile of ju-
doka during ukemi. Ukemi in a literary is the technique that must incorporate the
least chance for injury during falling. When judoist falls to the mat, the more area
of impact it has, the less damage it receives.
Keywords: acceleration measurement, wearable sensors, judo
25.1 Introduction
Judokas may suffer injury to any part of the body during a match (fight) and prac-
tices [9, 13]. In this kind of sports, regardless of male or female events, the most
performance ends with falling. When a man falls, the body hits the tatami and the
one that takes the impact. Putting a hand out to break a fall is a natural reaction. If
he or she put his hand down to stop the impact, the body weight lands in an area
about the size of palm and hits the mat first. That’s a lot of kilograms per square
centimeter impacting the hand, and that impact can damaging hand, wrist, elbow
and shoulder. It is caused by the shock wave of energy traveling through these parts
of human body. This human reaction comes from a fear of a head and torso hitting
against a mat. The analysis of kinematics of body movement during forced fall has
been examined in paper [10]. In this work, tested person was trotting in place on the
Technical University of Koszalin, 75-453 Koszalin, Poland

e-mail: sebastian.glowinski@tu.koszalin.pl
WWW home page: http://wtie.tu.koszalin.pl

https://doi.org/10.1007/978-3-319-70063-2_25
236 Sebastian Glowiński et al.
hard ground and was unbalanced by an assistant by pulling one of the legs. Tested
person was not aware in which direction would be knocked out by the force, causing
the fall. The analysis of kinematics of body during a collision with vertical obstacle
was conducted in the work [11]. The measurements have been conducted by using
MVN Biomech system with inertial sensors. The velocity and displacement values
of the body mass centre were presented. Whereas in this work, the authors are fo-
cused on forces, acting on parts of human body. The purpose of the current study
were: (1) to explore and analyze how is an acceleration of human body during typi-
cal Ukemi; and (2) to asses the generic risk of injury of human body during contact
with the mat.
25.2 Ukemi
According to Ohlencamp the most universal and basic of all fears, across all cul-
tures, is the fear of falling [15]. The most people can also overcome their fears and
learn to cope with them. Learning to deal with the fear of falling can establish an
important sense of confidence. For example in judo it is also not possible to be suc-
cessful in mat fighting without learning to control the transition from standing to the
tatami, to a position as a result of an opponent (Tori) action. There are four basic
breakfall in judo: Koho Ukemi, Yoko Ukemi, Mae Ukemi and Zenpo Kaiten Ukemi.
25.2.1 Koho Ukemi
Koho-Ukemi (ang: The fall back) is the controlled fall back, with the rolling on
the back Fig.25.1. During this fall, it is not allowed fall on the straight, flat back,
because it threatens very serious consequences for the health. It needs to be bent
strongly in the arch, in order to make possible rolling on the back from the bottom
to shoulders. It causes the spreading the strength of fall on rotational movement.
During the rolling the back, when shoulders are in the mat level, the both hands (at
the 45◦ angle) hit strongly the tatami in relation to the trunk (torso). In the middle of
koho ukemi the head is bent forward, so that the chin touches the chest. That position
of head prevents hitting on the mat, which threatens serious injuries and even death
[2, 3]. At the end body stops in the final position (the laying position on the tatami).
According to Koshida the judo backward breakfall skill may play an important role
in decreasing the number of head injuries [8]. In this case the essential element in
falling energy dissipation is arms action, which makes force reaction.
25 Body part accelerations evaluation for chosen techniques in martial arts 237
Fig. 25.1: Koho Ukemi (Rear falling)
25.2.2 Yoko Ukemi
Yoko Ukemi (ang: The fall aside) is the controlled fall aside, on the side part of the
back Fig.25.2. In first part of Yoko Ukemi judoist places chin on the chest to keep
the head from bouncing off the floor. It is very important, for example if the neck
is not restrained, judoist may very well gets a whiplash, as his body impacts the
mat and his head continues with inertia towards the floor, snapping the head in a
downward direction. Therefore, the next part of this Ukemi is an arm and leg swing
across judoist front. By the way, if he is falling left, he should swing the left arm to
the right. At the same time, he slides the leg on the same side as the swinging arm
into the other leg. During collapsing to the side one leg just swung for the moment.
When judoist falls further, whatever body parts hit the tatami. The torso is the one
that takes the impact. By slapping the tatami the energy of the impact disperses out
and away from judoist rather than getting absorbed by him. By slapping the tatami
simultaneously as his side, the kinetic energy is dissipated [4]. Again, also in this
case the essential element in falling energy dissipation is arms action, which makes
force reaction.
Fig. 25.2: Yoko Ukemi (The side breakfall)
25.2.3 Mae Ukemi
This technique fall helps to prevent judoist head from hitting the mat if he is attacked
from behind [14]. Judoist falls so that his forearms contact the tatami on landing
Fig.25.3. During falling hands should be turned inward about 45◦ , so that elbows
bent outward. Judoist should flap the mat hard while looking away to one side.
Contact with the tatami should be as flat as possible over entire forearms. In this case
the essential element in falling energy dissipation is forearms action, which makes
force reaction. Not enough forearms slap may cause the falling energy cumulation
in a judoist shoulders and successive injury.
Fig. 25.3: Mae Ukemi (Front or Forward Breakfall)
25.2.4 Zenpo Kaiten Ukemi
In this Ukemi judoist rolls over (to either his left or right side) in somersault action:
from the right-leg lead, he initiates a forward somersault action, tucking head in
while rolling over. A right-leg lead results in slapping judoist left hand and vice-
versa. Next, he contact the tatami simultaneously with his legs, hands and body,
landing on his side [12, 14]. As his body comes around to land, judoist synchronizes
legs and arms to hit the mat flat. He land on his side, feet apart, knees flexed, head
up, and slapping hand flat to the tatami, protecting head and spine. In this case
the essential element in falling energy dissipation are simultaneously arms and legs
action, which make force reaction. Additionally, body rotation and rolling on the
mat, convert and dissipate kinetic energy.
Fig. 25.4: Zenpo Kaiten Ukemi (Forward Roll)

25.3.1 Participant
One elite male judoist (age: 54 years, body height: 1.71 m, body mass: 75 kg) vol-
unteered to participate in the study. The subject was a member of Gwardia Koszalin
judo team and was healthy without any muscular or tendonitis injuries in body.
25.3.2 Wireless sensing system
The microelectromechanical system is composed of six body area wireless inertial

sensor nodes based on ProMove mini platform [16]. The system creates the bridge
between inertial measurements units and wireless sensor networks. It is embedding
in one device the following: 10 degrees-of-freedom inertial sensors: ±2 to 16 g
accelerometer, ±250 to 2000◦ /s gyroscope with resolution 0,007◦ /s and ±250◦ /s
range and magnetic field intensity (compass), 2 GB flash memory in each sensor,
low-power RF transceiver in the 2.4 GHz license-free band. Inertia gateway as a
central hub for synchronized data collection <100 ns. Battery life 4 h in full stream-
ing mode. Detailed parameters with description were presented in [6]. As shown
in Fig. 25.5, the six ProMove mini sensors are placed on the leg, torso, head and
the arm and forearm. For each sensor the orientation is calculated relative to Earth
reference frame in terms of roll, pitch and yaw angles.
5 6
4
3
Fig. 25.5: The human model with nodes placement

25.3.3 Experiment procedure
After being informed about the test procedure and the possible risks involved in the
study, the subject reviewed and signed a consent form for participating in this study.
The judoist first took fifteen minutes for warm-up, including light jogging, stretch-
ing and several easy judo elements. Then, nodes were placed on the body and the
gateway was connected to the computer. The sensor sampling was 200 Hz. Next the
sensors were turn on and verified that data was received from all of them. The start-
ing position was Shizen Hontai (with neither foot/leg leading) similar to an anatom-
ical position. The nodes were resetting and the algorithm calculated the orientation.
After resetting, it was necessary to hold the nodes still (without movement) for a
few seconds to stabilize their orientation. After starting recording, judoist repeated
five times each Ukemi in order to compare the results whether they are similar in
value. Next, the all raw data from each sensor were send to a central computer via
the USB. Angles, angular velocity were measured but only accelerations were cho-
sen for biomechanical analysis in this study. It is assumed that the expected results
can be generalized regardless of only one judoist was tested.
25.4 Results
The results of experiments are presented in Fig. 25.6 to 25.9.The kinetic energy of
an object is its property due to its motion. It is defined as the work needed to ac-
celerate a body of a given mass. Therefore, the acceleration is directly related to its
kinetic energy. In case of each Ukemi, there are two sets, which underline an action
of essential for the movement and energy dissipation body parts i.e. arms and legs.
The technique needs the different time to do it properly. Each of them starts with a
specific initial movement. In the figures, there are the time periods shown, which are
related to the phase of Ukemi presented in Fig. 25.6 to 25.9 (without initial move-
ments). The Koho Ukemi protects the rear part of human head and back the strong
strokes. As it is shown in Fig. 25.6 the energy of body rear falling is dissipating
by strong ams hit (acceleration over 250 m/s2 ) against Tatami with supporting legs
movement with acceleration more then 40 m/s2 . The maximal acceleration of head
and torso is about 50 m/s2 . The Yoko protects the side part of head and body the
strong stroke. Here, as it is shown in Fig. 25.7 also the arms strong stroke (acceler-
ation over 250 m/s2 ) dissipates the judoist falling energy. Because of only one hand
acts on tatami the legs action is stronger (acceleration nearly 80 m/s2 ). In this case
both head and torso reach higher acceleration (head 60 m/s2 - torso - 75 m/s2 ).
The Mae protects front of human body the stroke in tatami. In this case seems
only arms play a main rule. The sensors on the legs recorded significantly less ac-
celeration. As it is shown in Fig. 25.8, in this case both head and torso reach also
high acceleration (head 80 m/s2 - torso - 70 m/s2 ).
The specific Ukemi is Zempo Kaiten, when the rotational movement of the body
is applied. In case of this kind of technique, in contrast to Koho Ukem, the rotation
arms action
body part acceleration [m/s 2]

300
sensor 4-forearm
sensor 5-upper arm
sensor 3-torso
200
sensor 6-head
100
0
5 5.5 6 6.5 7 7.5
time [s]
legs action
60
sensor 1-shank
sensor 2-thigh
sensor 3-torso
40 sensor 6-head
20
0
5 5.5 6 6.5 7 7.5
time [s]
Fig. 25.6: Koho Ukemi (Rear falling) - accelerations of chosen body parts
arms action
300
sensor 4-forearm
sensor 5-upper arm
sensor 3-torso
200
sensor 6-head
100
0
2 2.5 3 3.5 4 4.5 5
time [s]
legs action
80
sensor 1-shank
sensor 2-thigh
60
sensor 3-torso
sensor 6-head
40
20
0
2 2.5 3 3.5 4 4.5 5
time [s]
Fig. 25.7: Yoko Ukemi (The side breakfall) - accelerations of chosen body parts
is not stopped and judoist eventually return to Shizen Hontai position. This feature
and additionally significantly stronger action of legs (acceleration more then 150
m/s2 ), as it is shown in Fig. 25.9, reduce the reaction on the head and torso to less
then 50 m/s2 .
arms action

300
sensor 4-forearm
sensor 5-upper arm
sensor 3-torso
200
sensor 6-head
100
0
11 11.5 12 12.5
time [s]
legs action
100
sensor 1-shank
80 sensor 2-thigh
sensor 3-torso
sensor 6-head
60
40
20
0
11 11.5 12 12.5
time [s]
Fig. 25.8: Mae Ukemi (Front or Forward Breakfall) - accelerations of chosen body
parts
arms action
300
sensor 4-forearm
sensor 5-upper arm
sensor 3-torso
200
sensor 6-head
100
0
2.5 3 3.5 4 4.5 5
time [s]
legs action
200
sensor 1-shank
sensor 2-thigh
150
sensor 3-torso
sensor 6-head
100
50
0
2.5 3 3.5 4 4.5 5
time [s]
Fig. 25.9: Zenpo Kaiten Ukemi (Forward Roll) - accelerations of chosen body parts
25.5 Discussion
There are not much works, which present results focused on forces, accelerations of
particular parts. The fall techniques during the training and fall skills in a reduction
of impact force during a fall are analysed [5, 10, 11]. Trained ability of safe fall is
the most effective prevention of body injuries caused by fall and collision with the
ground (mat) or obstacle should be expanded in the body parts action analyses. A
perfect model should combine, not only a body as one mass, but as system. There-
fore, the values of acceleration are underlined in the work. It is assumed that the
expected results can be generalized regardless of only one judoist was tested. The
chosen judo Ukemi data according to proper method, should give the similar results
in case of every experienced sportsman. The measurements shows significantly high
accelerations appear, when Ukemi technique was done. The highest were recorded
in case of arms (more then 250 m/s2 ) and legs (more then 150 m/s2 ) in one case
of Zempo Ukemi. It shows that Ukemi technique is based on a force reactions pro-
duce mainly by arms in order to reduce, nearly three times (50 m/s2 - 80 m/s2 ),
acceleration and in the same way force acting on a sensitive body parts i.e. head and
torso. On other hand the Ukemi technique can be risky in case of an unexperienced
person. Wrong arms or legs position especially during their action (stroke against a
tatami) may cause wrist, elbow, shoulder, ankle or knee injuries. Those inertia sen-
sor devices were in practical use already. Now, we need an innovative methodology
for the interpretation from the sensor time series to the human skill or knowledge.
References
1. James D.A.: The Application of Inertial Sensors in Elite Sports Monitoring. In: Moritz E.F.,
Haake S. (eds) The Engineering of Sport 6. Springer, New York, 289–294 (2006)
2. Ishikawa, T., Draeger, B.: Judo Training Methods. Tuttle Publishing, pp. 328, (1999)
3. Murat, R.: Karate For Beginners And Advanced. Atlantic Publishers and Distributors Pvt Ltd,
pp. 344, (2007)
4. Stevens, R., Semple, E.: The Fundamentals of Judo. The Crowood Press Ltd, pp. 263, (2014)
5. Groen, E., Smulders, E., de Kam, D.: Martial arts fall training to prevent hip fractures in the
elderly. Osteoporos Int, 21, 215–221 (2010)
6. Glowinski, S., Blazejewski, A., Krzyzynski, T.: Inertial sensors and wavelets analysis as a tool
for pathological gait identification. Advances in Intelligent Systems and Computing, Innova-
tions in Biomedical Engineering, 526, 106–114 (2016)
7. ProMove wireless inertial sensing platform. http://www.inertia-technology.com/promove-mini
8. Koshida, S., Ishii, T., Matsuda, T., Hashimoto, T.: Biomechanics of the judo backward breakfall:
comparison between experienced and novice judokas. Archives of Budo, Science of Martial
Arts, 10, 187–194, (2014)
9. Kujala, UM., Taimela, S., Antti-Poika, I., et al.: Acute injuries in soccer, ice hockey, volleyball,
basketball, judo, and karate: analysis of national registry data. Br Med J, 311, 1465–1468 (1995)
10. Michnik, R., Jurkojc, J., Wodarski, P., Mosler D., Kalina, RM.: Similarities and differences of
body control during professional, externally forced fall to the side performed by men aged 24
and 65 years. Archives of Budo, Science of Martial Arts, 10, 233–243 (2014)
11. Michnik R., Jurkoj? J., Wodarski P., Mosler D., Kalina R.,M.: Similarities and differences of
the body control during professional collision with a vertical obstacle of men aged 24 and 65.
Archives of Budo, Science of Martial Arts, 11, 27–39 (2015)
12. Otaki, T., Draeger, DF.: Judo Formal Techniques: A Complete Guide to Kodokan Randori no
Kata. Tuttle Publishing, pp. 456, (2011)
13. Pieter, W.: Martial arts injuries. Med Sport Sci., 48, 59–73, (2005)
14. Takahashi, M.: Mastering judo. Human Kinetics Publishers, pp. 225, (2005)
15. Ohlencamp, N.: Overcoming The Fear Of Falling. http://www.fightingarts.com/reading/article.php?id=134
(access: 01.05.2017)
16. www.inertia-technology.com (access: 20.04.2017)
Chapter 26
Determination of the number and frequency of
the steps for gait with elbow crutches based on a
crutch acceleration
Magdalena Długosz1 , Piotr Wodarski1 Andrzej Bieniek1 , Miłosz Chrzan1 , Marek

Gzik1 , Kamil Joszko1 , and Jarosław Derejczyk2
Abstract. Gait is one of the basic and most important forms of human locomo-
tion. Gait can be described by parameters such as: speed, step length, stepping rate,
ground reaction force values. One way to determine these parameters is analysis of
acceleration of selected body segments. There are not many algorithms in the lit-
erature that could be used to analyze gait with the elbow crutches. Accordingly, an
algorithm should be developed to determine, for example, the frequency and num-
ber of steps based on the crutch acceleration. In the following study, signal from
IMU sensor located on elbow crutch was used to determine the parameters. The
algorithm based on double filtration of recorded acceleration and peak detections
from obtained signal. Results were compared and correlated with MVN Biomech
motion capture system (high accuracy system). The discussion chapter considers
uncertainty of the measuring methodology and the possibilities of using the de-
signed system.
Keywords: human gait, crutches, acceleration, number of step determination
26.1 Introduction
Gait is one of the basic and the most important forms of human locomotion. Dys-
functions which are associated with lower limbs leads to impairment or complete
lack of locomotor function and may reduce mental and physical comfort [8]. As a
result, the patient is often withdrawn from social life.
Gait can be described by several parameters such as: speed, stride length, fre-
quency gait, ground reaction force values [6, 8, 9]. One way to determine the
aforementioned parameters is a analyze of accelerations of selected body segments.
Methodology of measurement of gait speed and step length requires complex of sig-
Silesian University of Technology, Faculty of Biomedical Engineering, Poland

e-mail: magdadlugosz93@gmail.com,piotr.wodarski@polsl.pl, · Katowice -
Szpital Geriatryczny im. Jana Pawła II

https://doi.org/10.1007/978-3-319-70063-2_26
246 Magdalena Dáugosz et al.
nal processing algorithms. First step is filtration of acceleration signals from IMU
sensors because of noises, measuring range limits and interferentions among of parts
of components inside sensor. In [11] authors present the using of the quaternion
Kalman filter in real-time tracking of human body movement. The Kalman filter
processes data from triaxial angle sensors, accelerometers and magnetometers. Fil-
tration of measured acceleration allowed for their double integration. The result was
a good approximation of the path walked upright by the investigated subject.
It is also possible to develop a system for studying and analyzing gait based on
accelerometer data from IMU sensors. The authors [7] placed an accelerometer on
the patient’s body during a walk and the results of the measurements allowed to
determine: average gait cycles, mean acceleration in one cycle, percentages of each
gait, and acceleration ranges. Acceleration measurements were made using triax-
ial acceleration sensors, located on various anatomical points of the human body:
the ankle, the knee, the hips, the cross and cervical spine and the head. A similar re-
search methodology was used in [2], where vertical acceleration signals recorded by
the accelerometer placed on the ankle of the examined subject were used to detect
particular gait cycles. However, in reference [12] gait parameters determined based
on the acceleration obtained from the accelerometer located on the human lower
back. The steps length and average walking speed were determined by the torso
movements up and down. The center of gravity in the sagittal plane was traversed
by circular trajectories during a single walking cycle - the so-called inverted pendu-
lum model. The height changes in the center of mass in this model depended on the
length of the step. This algorithm was extended in the work [1], where the authors
added one parameter, which is the constant determined by the calibration for each
test subject. While at work [4] was designed portable wireless sensor network used
to do real-time monitoring kinematics during gait. For (in this case) lower limb ac-
celeration analysis applied wavelets tranform. This system may allow the detection
of specific gait features and their possible abnormalities .
Literature review of solutions for gait parameters determination based on ac-
celerometers does not provide unequivocal information according to correct location
of the IMU sensors [9]. The appropriate location of the acceleration sensor would al-
low to obtain measurements of parameters such as step length or frequency of steps
with the most accuracy. Placing a sensor on the human body may cause discom-
fort in movement. especially if the person has a dysfunction of the motor system.
In this case it is possible to place the sensor on the elbow crutches or the balcony
with which the patients move. There are no algorithms in the literature that could be
used to analyze people walking with the elbow crutch. For example, the authors [10]
use only force sensors located on an orthopedic crutch to estimate skeletal-muscular
load of the upper limb. In addition, an optical system with passive markers was used
to measure temporal spatial parameters, both on the body of the subject and on the
crutches. In turn, in the work [3] optical system was also used to determine temporal
spatial parameters.
Literature review indicates needs of new methodology dedicated for example to
determine the frequency of steps based on the acceleration of the elbow crutch with
26 Number and frequency of the steps 247
IMU sensors. In this scope of study, it is possible to modify existing algorithms used
for measuring the accelerations of sensors placed on the human body.
The aim of the study is determining of methodology of calculation of averaged
frequency of steps and number of steps during gait with the elbow crutch on the
basis of the acceleration measurements from IMU sensor placed at the elbow.
26.2 Methodology
Detection of gait parameters based on acceleration of el-

bow crutch requires the location of the accelerometer and
the recording system on the elbow crutch and then valida-
tion of the recording data. For this purpose, the MPU 6050
(3-axis accelerometer and gyroscope) and one of the XSens
MVN Biomech [13] sensors were attached to the classic
elbow crutch, below the handgrip. For the first sensor, the
maximum gyro operating range is ±2000o /s, while the ac-
celerometer ±4g, for the MVN Biomech the range of gy-
roscopes is ±1200o /s while the accelerometer ±16g. The
first system recorded acceleration data with a constant fre-
quency of 9Hz while the second system with a frequency
of 120Hz. This is the highest sampling rate of this system. Fig. 26.1: The elbow
Data recorded by the first sensor was stored on the MicroSD crutch with mea-
card, while the reference measurement system transmitted suring sensors with
the measured and prefiltered values to the computer in wire- marked location of
less way. Systems calibrations were made according to the systems.
producer’s recommendations. The view of the study with the
prepared elbow crutch with sensors is shown in fig. 1.
In order to verify the correctness of the measured acceleration, 14 measurements
of the acceleration of the elbow crutch during the walk were made, each included
from 10 to 30 steps with elbow crutch. Finally 14 data files from the elbow crutch
and 14 data files recorded by the MVNBiomech system were received. The time
of data acquisition depends on speed of walk. In order to perform a comparative
analysis of acceleration it was necessary to interpolate data from the first system
to 120 samples per second (sampling rate of the second system is 120Hz). In the
next step resultant accelerations were calculated. The analysis was based on the
calculated RMS for each of the waveforms and the calculated Pearson correlation
coefficients between waveforms using the MatLab software.
After the initial verification of measurements, a second series of measurements
was started, the person walked with elbow crutch following number of steps: 10, 20,
and 30 (four times for each attempt). The recorded results were used to determine
parameters related to the number of stride and their frequency. Initially, the analysis
of the components of the obtained waveforms was performed by calculating the Fast
Fourier Transform (FFT) of resultant acceleration. The next step was the filtration
of the acceleration waveforms using two IIR digital filters. The first filter was a
low-pass filter with a 1 Hz cutoff frequency. This cutoff frequency was determined
by the assumption that average gait of the person with the elbow crutch would not
be faster than the average gait of the healthy person (frequency of the stride in a
healthy person is gait does not exceed 1Hz). The next filter was a high-pass filter
with a cutoff frequency of 0.1Hz that cut off the constants from the waveforms.
For both filters the damping in the band was at least 80dB. In the last step, for the
obtained after filtration waveforms, peaks detection were made by localizing local
maxima which occur no often than 1s (1Hz). Finally, it was assumed that the number
of found peaks and the results defined half of the steps taken by the person. The time
between successive detected peaks was half of the frequency of the steps. The entire
step detection algorithm can be represented by a graph as in Fig. 2.
Fig. 26.2: Algorithm of steps detection analysis.
In the last step, the average stride frequency was determined as the inversion of
the mean time value between successive detected stride for each recorded sample.
26.3 Results
Figure 26.3, on the left, shows an sample graph of the measured acceleration us-
ing the MVN Biomech system and the MPU6050 system, while on the right graph
shows the frequency components of these acceleration (it is a spectrum of the mea-
sured acceleration signal).
For the measured (using two systems) 14 resultant accelerations RMS and Pear-
son’s correlation coefficients were calculated. The results are shown in table 26.1.
Table 26.1: Root mean square (RMS) and Pearson correlation coefficient for regis-
tered data.
Sample number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 AVG SD
RMS 3,81 3,57 3,91 7,23 4,49 6,35 5,52 5,07 3,37 4,17 7,26 6,75 4,06 7,38 5,21 1,50
Pearson correlation coefficient 0,69 0,57 0,57 0,61 0,68 0,57 0,59 0,66 0,64 0,63 0,57 0,56 0,57 0,51 0,60 0,05
Fig. 26.3: Synchronized data from system MPU 6050 and MVN Biomech System
and data from frequency analysis.
Sample of results of the second study are shown in figures 26.4. The charts show
obtained resultant accelerations and their frequency analysis for 10, 20 and 30 steps
with the elbow crutch.
Fig. 26.4: Acceleration from MPU6050 and frequency analysis for A - 10 steps (5
double steps), B - 20 steps (10 double steps), C - 30 steps (15 double steps)
Table 26.2 shows the results of the stride (double steps) detection for the MPU
6050 and MVNBiomech systems for each test and the differences between the ob-
tained values.
Table 26.2: Double steps detection results for each study.

Sample num- Data from: Actual number Number of stride Diff. between sys- Diff. between acutal and de-
ber: of stride (double (detected): tems: tected numer of double steps:
steps):
MPU 6050 5 5 0
1 0
MVN Biomech System 5 5 0
MPU 6050 5 6 1
2 0
MPU 6050 5 5 0
3 0
MPU 6050 5 5 0
4 0
MPU 6050 10 9 1
5 1
MPU 6050 10 10 0
6 0
MPU 6050 10 10 0
7 0
MPU 6050 10 10 0
8 0
MPU 6050 15 19 4
9 0
MPU 6050 15 15 0
10 1
MPU 6050 15 15 0
11 0
MPU 6050 15 15 0
12 0
Table 26.3 shows the results of the mean steps frequency for all measurements
Table 26.3: Average frequency of double step for each study.

Sample number: 1 2 3 4 5 6
Data from: MPU MVN MPU MVN MPU MVN MPU MVN MPU MVN MPU MVN
Frequency of stride (double step) [Hz]: 0.59 0.58 0.62 0.56 0.54 0.54 0.61 0.61 0.49 0.55 0.60 0.62
Frequency of steps [Hz]: 1.18 1.16 1.23 1.12 1.08 1.08 1.21 1.21 0.98 1.11 1.21 1.23
Number of steps per minute: 71 70 74 67 65 65 73 73 59 66 72 74
Sample number: 7 8 9 10 11 12
Data from: MPU MVN MPU MVN MPU MVN MPU MVN MPU MVN MPU MVN
Frequency of stride (double step) [Hz]: 0.60 0.60 0.61 0.62 0.56 0.59 0.56 0.52 0.53 0.52 0.57 0.55
Frequency of steps [Hz]: 1.20 1.21 1.23 1.23 1.12 1.18 1.12 1.05 1.06 1.04 1.14 1.10
Number of steps per minute: 72 73 74 74 67 71 67 63 64 62 68 66
26.4 Discussion
The first stage of the conducted research allowed to verify the correctness of the
measured acceleration using two independent systems. The MVNBiomech system
with high accuracy of the test-confirmed measurements [5] and the system was used
as a benchmark. The second system, based on the MPU6050 sensor, is a low-cost
system developed for the purpose of research. Comparison of the measured values,
like in the publications [11], was carried out both using time values of acceleration
and using frequencies analysis of measured acceleration waveforms. Fig. 3 shows

examples of waveforms obtained using both accelerating systems. Visual compara-
tive analysis allows us to infer that the waveforms on the time side are much more
different than on the frequency side, it indicates a more significant analysis of tem-
poral differences. By analyzing the calculated for each of the 14 RMS measurements
between time waveforms it can be shown that most of them have approximate val-
ues. However, there are measurements where the differences are greater and it means
that differences between the recorded time acceleration values for both systems are
occured. Also, differences in the acceleration values at the initial moment of con-
tact of the crutch with the ground can also be seen in the time graphs. This may
be due to lack of lowpass filtering in the MPU5060 system, which occurs in the
MVN Biomech system. Analysis of the calculated correlation coefficients between
accelerations shows a strong correlation for all measurements. A strong correlation
means similarities of the shapes of the measured acceleration waveforms. The simi-
larity of the shapes of obtained acceleration waveforms allows to use the MPU6050
system to detect the number of steps and their frequency based on the proposed
methodology without analyzing the raw acceleration values.
In the second part of the study, in order to develop an algorithm for detecting
the amount of elbow crutch steps, four acceleration measurements recorded on the
crutch were made for known number of steps. For each of the waveforms obtained,
similarly like in articles [11], a frequency analysis was carried out. Examples of re-
sults the waveforms are given in figs. 4, 5 and 6 for successively 10, 20 and 30 steps
with the crutch. On each of the calculated frequency charts, significant harmonics
of approximately 0.5Hz can be observed. Maximum amplitude on a spectrogram in
this area may indicate the frequency of stride. This harmonic can be observed on
spectrograms for all measurements of 10, 20 and 30 steps. In the next step, similarly
to [11, 2], the measured waveforms were filtered. The frequency of steps higher
than the frequency in normal gait was filtered and the constant value of signal was
rejected. Filtration of the measured signals allowed to specify the frequency bands
from 0.1Hz to 1Hz. An exemplary acceleration time signal plot before and after fil-
tration is shown in figure 7. In the figure, one can see the acceleration waveform after
filtration and the number of peaks corresponds to the number of actually executed
double steps (corresponds to the number of taps with the crutch on the ground). For
each filtered acceleration, the number of peaks and the time between the peaks were
calculated. The results are shown in table 2. The number of stride (double steps)
actually taken is two times more then the number of crutch steps on the ground.
Table 2 shows the differences between the actual number of double steps and the
number obtained from the algorithm developed. For 4 from 12 cases it can be ob-
served differences between these values. To identify the cause of the differences, a
step detection was performed using MVNBiomech measurements, where the actual
number of double steps was identified by using of photo cameras and the synchro-
nization of the measured signal. The results are shown in table 2. The biggest dif-
ference is 4 steps, and after analyzing the sequence of frames from the cameras, it
turned out that it was the result of additional tapes with the crutch on the ground at
the moment of rotation during the crutch walk. However, the most disturbing was
the difference between the detection of the number of steps based on measurements
from the MPU6050 system and the MVN Biomech system such as for example no.
5. This difference may be due to an algorithm’s malfunction. However, it should
be pointed out that the difference in measurements between systems is a maximum
of 1 double step which can be assumed as a measurement uncertainty for the used
methodology.
In the last step of the study, the average frequencies of the double steps in each
measurement were calculated. The results are shown in table 3. The values obtained
are between 0.5Hz and 0.6Hz. This confirms the earlier assumption of the harmonic
detection of double steps with elbow crutch steps in the area of about 0.5Hz for
recorded waveforms. The frequency of single steps is twice the frequency value for
double steps.
26.5 Conclusions
The proposed methodology for the study and measurement of the elbow crutch ac-
celerations, based on two independent measuring systems, allowed the verification
of the obtained data on the shape of resultant acceleration trajectories. Measured
using a cheaper, developed system, the elbow crutch acceleration can be used to
estimate the average steps frequency and to estimate the number of steps taken by
a person moving with the elbow crutch. The proposed method of determining the
number of steps, based on double filtration and peak detection, can be easily imple-
mented into inexpensive microprocessor devices that can be permanently integrated
into the intelligent elbow crutch.
References
1. Alvarez, D., Gonzalez, R.C., Lopez, A., Alvarez, J.C.: Comparison of Step Length Estimators
from Weareable Accelerometer Devices. EMBS Annual International Conference 2006
2. Gafurov, D., Helkala, K., Srndrol, T.: Biometric Gait Authentication Using Accelerometer
Sensor. Journal of Computers, VOL. 1, NO. 7, October/November 2006
3. Gil-Agudo, A., Pérez-Rizo, E., Del Ama-Espinosa, A., Crespo-Ruiz, B., Pérez-Nombela, S.,
Sánchez-Ramos, A.: Comparative biomechanical gait analysis of patients with central cord
syndrome walking with one crutch and two crutches. Clinical Biomechanics 24 (2009) 551-
557
4. Glowinski, et. al. Inertial sensors and wavelets analysis as a tool for pathological gait iden-
tification, Springer, Innovations in Biomedical Engineering, Advances in Intelligent Systems
and Computing, DOI 10.1007/978-3-319-47154-9_13, 2016
5. Jun-Tian Zhang, Alison C Novak, Brenda Brouwer, Qingguo Li: Concurrent validation of
Xsens MVN measurement of lower limb joint angular kinematics, Physiological Measure-
ment, Volume 34, Number 8, 2013
6. Levine, D., Richards, J., Whittle, M.H.: Whittle’s Gait Analysis. Elsevier Urban & Partner
2014, chapter 2
7. Maj, P., Barczewska, K., Drozd, A., Kowalski, J.: Akcelerometryczny system badania i analizy
chodu [Gait analysis system based on accelerometers]. PAK vol. 58, nr 4/2012
8. Nowakowska K., Michnik R., Jochymczyk-Woźniak K., Jurkojć J., Mandera M., Kopyta I.:
Application of gait index assessment to monitor the treatment progress in patients with cere-
bral palsy (in:) Information Technologies in Medicine 5th International Conference, ITIB,
Kamień Ślaski,
˛ Poland, June 20-22, 2016 Proceedings, Advances in Intelligent System and
Computing 472
9. Nowakowska K., Michnik R., Jochymczyk-Woźniak K., Jurkojć J., Kopyta I.: Evaluation of
locomotor function in patients with CP based on muscle length changes. (in:) Innovation in
biomedical engineering, Gzik M., Tkacz E., Paszenda Z., Pi˛etka E.(eds.) Advances in Intelli-
gent System and Computing, vol. 526, p. 161-168
10. Requejo, P.S., Wahl, D.P., Bontrager, E.L., Newsam, C.J., Gronley, J.K., Mulroy, S.J., Perry,
J.: Upper extremity kinetics during Lofstrand crutch-assisted gait. Medical Engineering &
Physics 27 (2005) 19-29
11. Yun, X., Bachmann, E.R.: Design, Implementation, and Experimental Results of a Quaternion-
Based Kalman Filter for Human Body Motion Tracking. IEEE Transactions on robotics, VOL.
22, NO. 6, December 2006
12. Zijlstra, W., Hof, A.L.: Assessment of spatio-temporal gait parameters from trunk accelera-
tions during human walking. Elsevier 2002
13. https://www.xsens.com/products/mvn-biomech/ 29.06.17
Chapter 27
Factors influencing on mechanical properties of
porcine skin obtained in tensile test-preliminary
studies
Aneta Liber-Kneć and Sylwia Łagan
Abstract. Skin tissue shows sensitivity to many factors resulting from its biological
nature and also connected with mechanical tests conditions. The aim of the study
was to analyze the influence of sex and age of the animal and conditions of samples
storage on mechanical properties of porcine skin. The uniaxial tensile tests were
carried out for samples taken from two different pigs and stored before the test.
The basic mechanical parameters were compared. The study revealed sensitivity of
the skin samples to different factors. Comparison of skin samples taken from the
porker the and sow differing in age and weight showed that better substitute for the
human skin for mechanical properties tested in vitro is the porker skin. The use of
different storage variants exhibited similar influence of freezing and incubating in
saline solution for 2 days for porker samples, and significant influence of freezing
for 5 days.
Keywords: porcine skin, tensile test, storing conditions, tensile strength
27.1 Intoduction
The skin fulfils a number of functions important for the body. In the histological
structure of the skin, three main layers can be distinguished: epidermis, dermis, and
hypodermis. Dermis layer plays the most important role in terms of mechanical
properties [13]. Skin is an anisotropic, viscoelastic and non-linear material [10].
Results of experimental studies show that mechanical parameters of skin differ. The
value of tensile strength for human skin is reported between 2 and 28 MPa, the value
of modulus of elasticity ranges from 12 to 112 MPa [10], [13]. These differences
result from the biological variety among subjects, the sensitivity of biological tissues
Cracow University of Technology, Warszawska 24, 31-155 Cracow, Poland,

e-mail: aliber@pk.edu.pl,
WWW home page: http://pk.edu.pl

https://doi.org/10.1007/978-3-319-70063-2_27
256 Aneta Liber-Kneć et al.
to sample test and storage conditions, problems with obtaining samples of identical
dimensions (e.g. various thicknesses) as well as from the anisotropic character of
skin [6]. Also factors like age, body zone and hydration influence on the thickness
of each skin layers and on its properties [12]. It should be taken into account that
the properties of skin changing once removed from the body, the different testing
methods are employed and there is a lack of established boundary conditions of tests
[10].
Ethical issues and necessity of ensuring high homogeneity of the material (an
uniform of feeding, similar impact of external factors, the same age) affect on using
animal models in in vitro mechanical tests of the skin. Different animals like rat,
mouse and rabbit [9] have been used in experimental studies but porcine skin has
shown to be the most similar to human skin [14]. Compared to other laboratory
animals, pigs contain dermal collagen and elastic content that is more similar to
humans. Also physical and molecular responses of pigs to various growth factors
are similar to humans [14].
This study aims to capture the differences in material data obtained in tensile test
for skin samples taken from pigs differing in sex and age, and stored in different
conditions before the test.
27.2 Material and Methodology
Skin samples taken parallel to the back bone from the back of two domestic pigs
were used in this investigation. The first pig was 8-month old male porker, weighting
about 120 kg. The second pig was 2 years old female sow, weighting 220 kg. All
samples had the same dimensions: the length 100 mm and the width 10 mm, but
different thicknesses. The average thickness for the porker was equal to 2.50 ± 0.15
mm, and for the sow was 3.08 ± 0.22 mm. Samples were stored in the saline solution
(0.9 %) at the temperature of 4 ◦ C no longer than 12 hours (fresh) before the test.
Then, three variants of sample storing were applied before conducting the tests:
- in the saline solution (0.9 %) at the temperature of 4 ◦ C for 2 days (NaCl),
- frozen in polypropylene foil at the temperature of - 18 ◦ C for 2 days (frozen 2),
- frozen in polypropylene foil at the temperature of - 18 ◦ C for 5 days (frozen 5).
In case of frozen samples, tests were carried out after one hour from defrosting.
For the such prepared samples, the mechanical properties under uniaxial static
tension were determined with the use of the MTS Insight 50 testing machine. The
samples were mounted using scissor action grips with self-tightening (Fig. 1) and
they were extended at the speed of 5 mm/minute at room temperature. The measure-
ment base of the sample was 50 mm.
Each set of samples for tension testing (the samples divided according to the kind
of pig and the time of conducting the test from the moment when the samples were
taken) contained minimum 5 samples. Registered force elongation curves (force (F)
– elongation (Δ l)) were recalculated into stress (σ ) – strain (ε ) curves. Basic skin
mechanical properties, i.e. the tensile strength (σmax ), Young’s modulus (E), and the
Fig. 27.1: The method of fixing samples in the machine handles.
strain for a maximum force (ε ) were determined. For the calculations, a rectangular
cross section of the sample was taken. The value of Young’s modulus was calculated
as a ratio of the increase in the stress to strain value in the middle–rectilinear curve
(E) in which loads are transported by collagen fibres [3]. The value of the tensile
strength (σmax ) was determined on the basis of the maximum force measurement in
the reference to the cross–section of sample, whereas the strain (ε ) was determined
on the basis of the measurement of the extension which corresponds to the maximum
force in the reference to the sample measurement base. The determined values of
strength parameters were shown as the average values with a standard deviation (X
± SD).
The exemplary stress-strain curves for poker and sow samples were compared in
Fig. 2. These curves showed typical for skin tissue non-linear characteristic with
three phases [7]. In the initial loading phase, great deformations of the skin occurred
at a relatively low applied load. In this phase, the elastin fibres (which keep the skin
smooth) are mainly responsible for the stretching mechanism [7], [10]. The stress-
strain relation was approximately linear. In the second phase of stretching, load were
higher and the tissue became stiff at higher stresses. This phenomena is connected
with orientation of collagen fibres in the direction in which the load is applied [7],
[10]. In the third phase, the ultimate tensile strength was reached and the fibres
began to break.
Fig. 27.2: The exemplary stress-strain curves for porker and sow, fresh specimens.
Comparison of porker and sow samples showed differences in values of charac-

teristic parameters calculated on the base of elongation curves (Fig. 3 to 5). Samples
from younger pig’s skin (the porker) showed a value of modulus of elasticity higher
by 52% than samples from older one (the sow) (Fig. 3). The difference in values
Fig. 27.3: The average values of Young’s modulus for porker and sow under different
conditions of storage.
of tensile strength between tested samples from two pigs is much smaller and it is
within the limits of measurement error (Fig. 4). The value of strain corresponding
to the value of maximum force is higher by 53% for the sow skin compared to the
porker skin (Fig. 5).
The lengthening of the time of skin storage in the saline solution up to 2 days
resulted in decreasing the tensile strength for the porker samples and increasing the
tensile strength for the sow samples. Freezing the samples also caused the decrease
in the maximum stress for the porker samples and the increase in the maximum
stress for the sow samples. The largest decrease of 20% was noticed for the porker
samples frozen for 2 days. The largest increase of 26% was noticed for the sow
samples frozen for 5 days (Fig. 4).
Fig. 27.4: The average values of tensile strength for porker and sow under different
conditions of storage.
Analogically to the changes of the value of tensile strength, the largest decrease
in Young’s modulus was noted after freezing the porker samples for 2 days (by about
45%). For the sow samples the increase of the Young’s modulus value by about 26%
was observed after freezing for 5 days (Fig. 3).
The highest values of strain corresponding to the maximum force were noticed
for all variants of sample storing for the porker skin. While for the sow skin samples,
the strain values corresponding to the maximum force decreased for all variants of
sample storing. The lengthening of the sample storing time in the saline solution
influenced on the decrease by 20% in the strain value of the examined porker skin.
While freezing by 5 days caused the decrease of 60% in the strain value correspond-
ing to the maximum force. For the sow skin samples, freezing by 5 days caused the
increase of about 33% in the strain value corresponding to the maximum force (Fig.
5).
Fig. 27.5: The average values of strain at maximum force for porker and sow under
different conditions of storage.
While comparing the obtained test results to both the results for a pig’s skin
and for the human’s skin, great differences in the determined values of the strength
parameters can be noticed. The obtained values of strength and modulus of elasticity
for the skin samples taken along the pig’s spine differ significantly from the values
obtained for the pig’s skin by Żak et al. [16]. The uniaxial tensile behavior of the
pig’s skin obtained in this study agrees well with that of the human’s skin. The
stress versus strain curve up to maximal force is J-shaped, although slope of the
curve may differ from species to species. At the same time, the obtained ranges of
strength parameter values of the pig’s skin correspond with the tensile strength and
Young’s modulus values of the human’s skin determined by Ni Annaidh et al. [10].
The lengthening of the storage time influenced on changes of measured parame-
ters. The samples refrigerator storage in the saline solution caused the degenerative
changes of the tensile strength for the porker skin specimens. In general, saline so-
lutions are considered to be worse media than e.g. PBS, HHBSS, McCoy’s 5A [2],
[15]. Studies of preservation for skin grafts in PBS at 4◦ C showed that viability was
retained for 24 h [5], in case of McCoy’s 5A medium viability was preserved for 4
weeks [2]. Storage of specimens in the freeze state caused decrease of the tensile
strength and Young’s modulus only for the porker specimens. In our tests the freez-
ing temperature (amounted about -18 ◦ C) was sufficient for the short time storage,
but in case of long term storage of the skin, cryopreservation is the most commonly
used method. The skin can be stored in such conditions for months to years [8].
27.4 Conclusions
From the point of view of an analogy between the human skin and a pig’s skin,
mathematical models of a tissue growth process or tissue regeneration are worked
out in mechanics [4], [11]. In order to achieve this goal both quantitative and qual-
itative experimental research is necessary. Comparison of skin samples taken from
two kinds of pigs (the porker and the sow) differing in age and weight showed that
better substitute for the human skin for mechanical properties tests in vitro is porker
skin. Among other it is connected with age–associated changes in skin and changes
of collagen quality, affecting the tensile strength.
Completing the material for experimentation more than once requires some
amount of time to form an adequate research group. Thus, ways of sample stor-
age until the research is completed are searched for. The results of this work allow
for the determination of the influence of the skin storage manner on its mechanical
properties. It is important information from the point of view of clinical research,
in which a pig’s skin is used as a substitute for the human skin. The use of differ-
ent storage variants exhibited similar inuence of freezing and incubating in saline
solution for 2 days for the porker samples, and signicant inuence of freezing
for 5 days. This research allows for the determination of the goal functions for the
optimization of the storage conditions of the human skin necessary to form skin
grafts used in curing burns or vast wounds. The presented research determines the
directions of further research within the scope of the diversication of storage pa-
rameters, i.e. time, media (preservation liquids), and temperature. The methods, the
spots from where the skin was taken, and the preparation of the researched material
inuence on the obtained results [1], [3], [4].
Acknowledgements
The work was realized due to statutory activities M-1/6/2017/DS.
References
1. Adull Manan, N.F., Muhammad, J., Ismail, M.H.: Quantifying the biomechanical properties of
bovine skin under uniaxial tension. Journal of Medical and Bioengineering. 2(1), 45–48 (2013)
2. Ben Bassat, H.: Performance and safety of skin allografts. Clinical Dermatology. 23(4), 365–
375 (2005)
3. Corr, D.T., Hart, D.A.: Biomechanics of scar tissue and uninjured skin. Advances in Wound
Care. 2(2), 37–43 (2013)
4. Corr, D.T., Gallant-Behm, C.L., Shrive, N.G., Hart, D.A.: Biomechanical behavior of scar tissue
and uninjured skin in a porcine model. Wound Repair and Regeneration. 17, 250–259 (2009)
5. De, A., Mathur, M., Gore, M.A.: Viability of cadaver skin grafts stored in skin bank at two
different temperatures. Indian Journal of Medical Research. 128(6), 769–771 (2008)
6. Geerligs, M.: Skin layer mechanics. Ph.D. Thesis, Technische Universiteit Eidhoven., 27–30
(2010)
7. Groves, R.B., Coulman, S.A., Birchall, J.C., Evans, S.L.: An anisotropic, hyperelastic model
for skin: Experimental measurements, finite element modeling and identification of parameters
for human and murine skin. Journal of the Mechanical Behavior of Biomedical Materials. 18,
167–180 (2013)
8. Han, B., Bischof, J.C.: Engineering Challenges in Tissue Preservation. Cell Preservation
Technology. 2(2), 91–112 (2004)
9. Karimi, A., Rahmatic, S.M., Navidbakhsh, M.: Mechanical characterization of the rat and mice
skin tissues using histostructural and uniaxial data. Bioengineered. 6(3), 153–160 (2015)
10. Ni Annaidh, A.N., Bruyere, K., Destrade, M., Gilchrist, M.D., Maurini, C., Ottenio, M., Sac-
comandi, G.: Automated estimation of collagen fiber dispersion in the dermis and its contribu-
tion to the anisotropic behavior of skin. Annals of Biomedical Engineering. 40(8), 1666–1678
(2012)
11. Ogawa, R., Kazuhisa, O., Tokumura, F., Mori, K., Ohmori, Y., Huang, C., Hyakusoku, H.,
Akaishi, S.: The relationship between stretching/contraction and pathologic scarring: The role
of mechanical forces in keliod generation. Wound Repair and Regeneration. 20, 149–157 (2012)
12. Pailler-Mattei, C., Beca, S., Zahouani, H.: In vivo measurements of the elastic mechanical
properties of human skin by indentation tests. Medical Engineering and Physics. 30, 599–606
(2008)
13. Silver, F.H., Freeman, J.W., DeVore, D.: Viscoelastic properties of human skin and processed
dermis. Skin Research and Technology. 7(1), 18–23 (2001)
14. Swindle, M.M., Makin, A., Herron, A.J., Clubb Jr., F.J., Frazier, K.S.: Swine as models in
biomedical research and toxicology testing.Veterinary Pathology. 49(2), 344–356 (2012)
15. Zieger, M.A. J., Tredget, E.E., McGann, L.E.: Mechanisms of cryoinjury and cryoprotection
in split-thickness skin. Cryobiology. 33(3), 376–389 (1996)
16. Żak, M., Kuropka, P., Kobielarz, M., Dudek, A., Kaleta-Kuratewicz, K., Szotek, S.: Determi-
nation of the mechanical properties of the skin of pig fetuses with respect to its structure. Acta
of Bioengineering and Biomechanics. 13 (2), 37–43 (2011)
Chapter 28
Flow of subretinal liquid through the retinal hole
after surgery – mechanical model and FEM
simulations
Tomasz Walczak1 , Paweł Fritzkowski1 , Marcin Stopa2 , and Martyna Michałowska1
Abstract. The objective of this study is to analyze subretinal liquid dynamics by

means of numerical simulations in patients with retinal detachment after insertion
of intussusception. The main parameter being analyzed is the outflow velocity of
subretinal fluid through the hole in the retina. All calculations were performed us-
ing FEM and the results were evaluated in relation to the velocity of outflow when
considering parameters such as; i) speed of eyeball movement, ii) shift of intus-
susception in relation to the center of the hole in the retina, iii) viscosity of the
subretinal fluid. Results obtained suggest an optimal speed of eyeball movement for
patients to achieve maximal speed of outflow of subretinal fluid. This is a significant
factor in achieving improved clinical outcomes for patients.
Keywords: retinal detachment, CFD simulations, FEM analysis
28.1 Introduction
Detachment of retina describes the condition where the neurosensory retina is dis-
placed or torn by the subretinal fluid from the lower RPE layer. This condition typ-
ically results in impairment or loss of vision. The risk of detachment increases in
patients over 50 years of age. Aging may introduce changes to the structure of both
vitreous and retina. Vitreous becomes disconnected from the retina if the mechani-
cal parameters of retina fibers have reduced elasticity. Additionally, a tear may form
Faculty of Mechanical Engineering and Management,

Poznań University of Technology
Jana Pawła II 24, 60-965 Poznań, Poland
e-mail: tomasz.walczak@put.poznan.pl · Clinical Eye Unit and Pediatric Ophthalmol-
ogy Service,
Heliodor Świ˛ecicki University Hospital, Poznań University of Medical Sciences,
Przybyszewskiego 49, 60-355, Poznań, Poland

https://doi.org/10.1007/978-3-319-70063-2_28
264 Tomasz Walczak et al.
between the retina and the choroid. Conventional surgical treatment is to implant in-
tussusception into the sclera directly below the hole of retina [1, 3, 9, 10]. The result
of the implantation is an outflow of subretinal fluid into the eyeball and reattachment
of the retina to the RPE. Where retina fibers have been torn from vitreous they may
completely reattach to the retina. Clinical outcomes are typically successful with
recovery of vision in the treated eye.
It has been found that, apart from the many individual characteristics and type
of retinal damage, clinical outcomes depend significantly on the location of the
intussusception. Further, there is significant correlation between restoration of vi-
sion and patient compliance with physician instruction regarding post-surgical eye
movement exercise. The velocity of the eyeball movement determines the veloc-
ity of liquid flow within the eyeball that in turn impacts the velocity of the liquid
passing through the hole in the retina. Finally, gravitational force may be significant
therefore patients may be positioned for optimal flow through the retinal hole.
In this paper FEM numerical simulations [4, 11, 12] were used to investigate
the effect of different parameters on the rate of outflow of subretinal fluid through
the retinal hole, which significantly influences the time of the entire fluid flow and
consequently the patient’s recovery time after surgical treatment.
28.2 Numerical simulations
The main parameter to be analyzed was the rate of outflow velocity of the retinal
fluid through the hole in the retina and the assessment influenced the value of its key
parameters. The effect of eyeball movement velocity, which is related to postopera-
tive recommendations, has been studied more precisely. Often, patients are advised
not to use eyeballs to move the fluid out of the retina through the hole. Another im-
portant parameter is the displacement of the intussusception relative to the center of
the retinal hole. Rarely is this implant located directly below the hole and is usually
offset against the axis of the hole. Therefore, the effect of this shift on the parame-
ters of subretinal fluid flow has also been studied. In addition, the influence of the
viscosity of this liquid on the results obtained has been investigated, because in the
literature [5, 6, 8] various values are often given, the effect of this parameter on the
results obtained has been studied.
Numerical simulations of flow were performed using COMSOL Multiphysics. A
mechanical model based on the mutual interaction of viscous fluid (hypotonic fluid)
and hyperelastic body (retina) has been established [2, 6, 7]. Material parameters
and simulation parameters are presented in Tab 28.1.
In the simulations, a steady-state subretinal fluid flow (left to right) was assumed,
with boundary conditions (given velocity) at the left edge of considered region. The
speed was determined from the assumed eye movement with angular velocity given
by the formula:
Ω = ρ ∗ Ω0 ,
Table 28.1: Main simulations parameters.
Total number of elements 4868

Number of triangular elements 3704
Number of quadrangular elements 1164
Number of degrees of freedom 21513
Retina: E = 103 Pa
ν = 0.499
ρ = 1100kg/m3
Liquid: Dynamic viscosity η = 65710−6 Pa · s

ρ = 999kg/m
(water in 37◦C)
Basic speed of eyeballs Ω0 = 100deg/s
where ρ is the speed regulator (ρ = 0.2 for eye movement during reading). The
diameter of the spherical intussusception was 4mm and the diameter of the retinal
hole was 2mm. Retinal displacement and fluid velocity distribution were studied. In
addition, the average velocity through the retinal hole was determined.
In Fig. 1. the distribution of vertical component of velocity of the fluid in con-
sidered region is presented. The visible thin layer represents retina with a hole, and
the surrounding fluid flows from left to right. The speed of the outflow through the
hole is relatively small (maximum value is 0.0159 [m/s]). To increase it and make
possible to stick the retina to the part of sclera, an intussusception is inserted. As
can be seen in Fig. 2, it leads to increase the velocity of outflow twice (to the max-
imum value of 0.0307 [m/s]). In the second case, the observed changes in velocity
distribution near the hole are more significant.
The first considered problem is the influence of the intussusception offset on
velocity of outflow. To specify the best pre- and post-operation conditions compu-
tations were performed for variable angular speed of eyeball. In Tab. 2. one can
find the maximum outflow velocities v at four different shifts of implant Δ x (neg-
ative value indicates shift to the left). Additionally, the corresponding values of ρ
parameter is presented.
As can be seen, the largest value of outflow velocity strongly depends on shift
value and is achieved at different speed of eyeball. One can find that the optimal
situation is when implant is shifted to the left. However it should be considered, that
calculations were conducted for one-sided flow (left to right). From this point of
view, the best situation is when the intussusception is placed just below the retina
hole (Δ x = 0).
Figure 3. illustrates the effect of the eyeball speed on retina vertical displacement.
Obviously, the displacement increases with growing angular speed. It is worth men-
tioning that in practice too large deflection of retina can lead to negative effect. If
Fig. 28.1: Distribution of vertical component of velocity [m/s] in considered region

for ρ = 0.2.
Fig. 28.2: Distribution of vertical component of velocity [m/s] in considered region

for ρ = 0.2 after insertion of intussusception.
Table 28.2: Influence of intussusception´s offset on outflow velocity in dependence

on eyeballs speed.
Δ x[mm] ρ v[m/s]
-1.5 1.35 0.0032

-1 1.35 0.0034
0 1.30 0.0038
1 1.27 0.0040
1.5 1.22 0.0042
retina is highly deflected the fluid can flow in the opposite direction and we also
observed much higher value of horizontal component of fluid velocity.
Fig. 28.3: Influence of eyeball speed on retina displacement where: ρ – parameter

of speed, Δ y[mm] – maximum retina deflection in vertical direction.
Similar analysis was performed for average outflow velocity. Figure 4. shows the
dependency of the velocity on the angular speed of eyeball. Unlike in the previ-
ous case, the characteristics has a maximum at ρ = 1.3. It means that increasingly
growing angular speed does not necessarily improve the outflow value and medical
conditions.
Fig. 28.4: Influence of eyeball speed on average outflow velocity where: ρ – param-
eter of speed, v[m/s] – average outflow velocity in vertical direction.
Another factor considered in the study is one of the main properties of the sub-
retinal fluid: viscosity. Usually researchers assume its value as for water [5]. It seems
obvious that this parameter can be slightly different in real conditions. In Fig. 5 one
can see three very similar characteristics corresponding to three levels of angular
speed. Note that as in the case of Ω , an auxiliary parameter μ was introduced which
indeed denotes the relative dynamic viscosity of fluid:
η = μ ∗ η0 ,
where η0 is the dynamic viscosity of water (see Tab. 1). The presented curves indi-
cate that the fluid viscosity plays a significant role in the fluid-retina behavior.
As the simulations results indicate all the considered parameters have important
influence on subretinal fluid behavior. The outflow average velocity depends on both
geometrical and physical factors: the position of intussusception, angular velocity
of eyeball, and viscosity of fluid. All numerical experiments were performed for
one selected geometry of the problem. In general shape of detached retina and its
hole could be different especially when one consider the problem in the full three-
dimensional case.
It should be noted that in the calculations the gravity force was neglected. Some
attempts were made by the authors to include this factor in the physical model, but
the results indicated that gravity influence on outflow is meaningless. It may mean
that the angular velocity of eyeballs has the highest impact on average outflow value.
Fig. 28.5: Influence of viscosity on average outflow velocity for three different eye-
ball speeds where: μ – parameter of viscosity (1.0 for water), v[m/s] – average
outflow velocity in vertical direction.
One can conclude that the appropriate speed of the eyeball movements is of greater
importance than the fluid flow due to gravity itself.
28.3 Remarks
Our results are in conflict with the routine recommendation to avoid eye movements
following conventional retinal detachment surgery. Based on the simulations we
have observed that eyeball movements have a beneficial effect on the flow rate of
the fluid through the retinal hole. We can conclude that the average speed of these
movements is of about 130 deg / s (about 6.5x the speed of eyeball movement during
reading). Both for smaller and faster movement of eyeballs, the speed of fluid flow
drops. In addition, at higher velocity of the eyeball significant retinal deformation
occurs, which may result in re-collection of fluid into the retina and, consequently,
a decrease in the rate of outflow.
We have also observed that the central position of the intussusception relative
to the retinal hole is optimal, that each shift reduces the outflow velocity and at
the same time increases the deformation of the retina in the vicinity of the hole. In
addition, the retinal fluid viscosity has a significant effect on retinal deformation
and behavior of the retinal fluid. However value of this parameter is difficult to
estimate and the water viscosity is most often assumed [5, 6]. For the value of the
flow rate and hence for the time needed to complete discharge of the subretinal
fluid, geometry of biomechanical system is of great importance, but this has not
been analyzed in this work.
Acknowledgements
The presented research results, were funded with grants for education allocated by
the Ministry of Science and Higher Education in Poland (02/21/DSPB/3493).
References
1. Angunawela, R. I., Azarbadegan, A., Aylward, G. W., Eames I.: Intraocular Fluid Dynamics
and Retinal Shear Stress after Vitrectomy and Gas Tamponade. Investigative Ophthalmology
and Visual Science. 52, No. 10, (2011)
2. Bottega, W. J., Bishay, P. L.: On the Mechanics of a Detaching Retina. Mathematical Medicine
and Biology. 30, 287–310 (2013)
3. Clemens, S., Kroll, P., Stein, E., Wagner, W., Wriggers, P.: Experimental Studies on the Disap-
pearance of Subretinal Fluid After Episcleral Buckling Procedures Without Drainage. Graefe´s
Arch. Clin. Exp. Ophthalmol. 225, 16–18 (1987)
4. Donea, J., Huerta, A.,: Finite Element Methods for Flow Problems. Chichester, (2003)
5. Foster, W. J.: Bilateral Patching in Retinal Detachment: Fluid Mechanics and Retinal ´Settling´.
Investigative Ophthalmology and Visual Science. 52, (2011)
6. Foster, W. J., Dowla, N., Joshi, S. Y., Nikolaou, M.: The Fluid Mechanics of Scleral Buckling
Surgery for the Repair of Retinal Detachment. Graefe´s Arch. Clin. Exp. Ophthalmol. 248(1),
31–36, (2010)
7. Kling, S., Marcos, S.: Finite-Element Modeling of Intrastromal Ring Segment Implantation into
a Hyperelastic Cornea. Investigative Ophthalmology and Visual Science. 54, (2013)
8. Lakawicz, J. M., Bottega, W. J.: An Analysis of the Mechanical Behaviour of a Detaching
Retina. Mathematical Medicine and Biology. 32, 137–161, (2015)
9. Seabag, J.: Vitreous in Health and Disease. Springer, New York, (2014)
10. Williamson, T. H.: Intraocular Surgery. Springer, Switzerland, (2016)
11. Zienkiewicz, O. C., Taylor, R. L.: The Finite Element Method. Vol. 1âĂŞ3, fifth ed.
Butterworth-Heinemann, Oxford, (2000)
12. Zienkiewicz, O. C., Taylor, R. L., Nithiarasu, P.: The Finite Element Method for Fluid Dy-
namics, seven ed. Butterworth-Heinemann, Oxford, (2014)
Chapter 29
Impact of Vessel Mechanical Properties on
Hemodynamic Parameters of Blood Flow
Wojciech Wolański1 , Bożena Gzik-Zroska2 , Kamil Joszko1 , Edyta Kawlewska1 ,

Marta Sobkowiak1 , Marek Gzik1 and Wojciech Kaspera3
Abstract. The main purpose of this work was the analysis of the impact of me-
chanical properties of blood vessels on the alteration of hemodynamic parameters
of blood. Within the framework of this research strength tests of blood vessels were
conducted, which aimed to determine mechanical properties by means of a static
tensile test and a system of Digital Image Correlation (μ DIC). The tests were per-
formed using static strength testing machine MTS Insight 2. Subsequently, on the
basis of computer tomography (CT) images, two numerical models of blood ves-
sels were formulated in the Ansys CFX software programme. The first model does
not take into consideration the changes of the artery geometry, while in the second
model the vessel walls were modelled as deformable elements having preset me-
chanical properties. The developed numerical models enabled the assessment of the
hemodynamic parameters as well as the state of stress and deformation of the walls
during the blood flow through the vessel.
Keywords: Computational fluid dynamics (CFD), blood flow, artery, mechanical

properties, mechanical tests.

Silesian University of Technology, ul. F. D. Roosevelta 40, 41-800 Zabrze, Poland,
e-mail: {wojciech.wolanski@polsl.pl,kamil.joszko@polsl.pl,edyta.
kawlewska@polsl.pl,marta.sobkowiak@polsl.pl,marek.gzik@polsl.pl},
· Department of Biomaterials and Medical Devices Engineering,
ul. F. D. Roosevelta 40, 41-800 Zabrze, Poland
e-mail: bozena.gzik-zroska@polsl.pl,
· Department of Neurosurgery, Medical University of Silesia,
Regional Hospital, Plac Medykow 1, 41-200 Sosnowiec
e-mail: wojciech$\_$kaspera@poczta.onet.pl

https://doi.org/10.1007/978-3-319-70063-2_29
272 Wojciech Wolański et al.
29.1 Introduction
There has been an increasing number of people suffering from blood circulation
problems lately. As epidemiological studies indicate, factors such as: low physi-
cal activity, stress, nicotinism, arterial hypertension contribute to the occurrence
of pathological changes in blood vessels. An early diagnosis and prompt correct
treatment considerably increase the chances of prolonging patients’ lives. In the
recent years there has been large-scale research conducted with the application of
computer modelling of the blood flow dynamics (Computational Fluid Dynamics -
CFD) in the models of blood vessels. This research has revealed that a high level
of the wall shear stress - WSS affecting the vascular nodes is the initiator of patho-
logical changes (Brian at al. 2007, Ferruzzi at al. 2011, Schulze-Bauer at al 2003).
Nowadays, numerical simulations are becoming a part of the diagnosing process
and preoperative planning (Gzik-Zroska at al. 2013, Wolański at al. 2017). The con-
ducted numerical simulations enable the determination of the distribution of stress
and strain during the blood flow in the artery, which may provide specialist doctors
with vital information (Humphrey at al. 2012, Lasheras 2007). With a view to the
above, this work aimed to develop a numerical model of a blood vessel and perform
simulation tests of the impact of the rigidity of the vessel wall on the value of the
intravascular pressure.
29.2 Methods
The analysis of the blood flow through blood vessels was conducted in a multi-
stage process. In the first stage, the properties of blood vessels were determined
using micro digital image correlation and strength testing machine MTS Insight 2
(Gzik-Zroska at al. 2016). Artery samples were prepared and subjected to a tensile
test after their ends had been fixed in special grips. The test was done in quasi-static
conditions at a velocity of 5 mm/min. Before the performance of the test the initial
distance of the grips and the thickness of the vessel were measured. Next, the mate-
rial was pulled in the direction of the longitudinal axis of the vessel. Experimental
studies have also included tests to determine the strains and deformations of speci-
mens during the angioplasty process. It was used the micro digital image correlation
system (μ DIC) that allows to observe the complete deformation of very small ob-
jects (17x17 mm and smaller). The system also enables to perform measurements
with very high accuracy in deformation and strain (0.01 pixel of deformation and
0.01% of strain). Apochromatic stereoscopic optical system is coupled with two
cameras, that continuously observe the object under test. Prior to the experiment the
random speckle patterns were applied to artery samples, that was necessary to ob-
serve the deformation of specimen in longitudinal and transversal direction during
the application of loads. In this study the load was generated by the balloon filled
with saline, on which the artery was embedded. The pressure in balloon were in-
creased using a standard device used in angioplasty procedure. It was applied the
29 Impact of Vessel Mechanical 273
pressure in the range of 0 to 10 bar with an increment of 1 bar. In μ DIC system it

was obtained the map of displacements in two axes, from which it was calculated
the Poisson ratio (mean value 0.36). The Young Modulus of vessels obtained during
experimental studies was 2,3 MPa (fig. 29.1). This parameter was subsequently used
to determine the mechanical properties in the modelling process.
Fig. 29.1: Displacement map obtained in μ DIC system
The next stage included the development of a geometric model of the vessel. To
build the model the Mimics software programme was used, which enabled segmen-
tation of two dimensional CT images and development of a three dimensional ge-
ometric model (fig. 2(a)). Subsequently, discretization of the model was conducted
(fig. 2(b)), which was then exported to the Ansys CFX software programme. The
blood flow analysis used the values of blood parameters defined by literature (Fung
1993, Holzapfel & Ogden 2010, Lasheras 2007): molar mass - 18,02 kg/kmol, den-
sity - 1050 kg/m3 , thermal capacity - 4181,7 J/kg∗K, viscosity - 0,0035 Pa∗s. Fig.
2(c) presents the view of the vessel model with marked boundary conditions. The
value of the blood flow velocity was obtained on the basis of the Doppler ultrasound
examination. The velocity value was preset on the inlet surface in the analyzed ves-
sel model - 0,98 m/s. The boundary conditions, which were defined in such a way,
made it possible to conduct the analysis of the blood flow at a temperature of 37o C
and reference pressure amounting to 1 bar.
(a) (b) (c)
Fig. 29.2: Model of the vesel: a) geometric model, b) discrete model, c) model with
marked boundary conditions
29.3 Results
Having completed numerical calculations, the following values were obtained: pres-
sure (fig. 3(a)), wall shear stress (WSS) on the surface of the vessel induced by the
flow and lines representing the flow velocity (fig. 3(b)). The results describing the
(a) (b)
(c)
Fig. 29.3: Strength analysis of the artery: a) applied pressure, b) equivalent (von-
Mises) stress, c) total deformation
distribution of pressure are especially important as they were used in the next step
to calculate deformations of the artery wall.
Mechanical properties of the artery obtained on the basis of experimental tests
(Young’s Modulus - 2,3 MPa, Poisson’s Ratio - 0,36) enabled the analysis of the
blood flow taking into consideration the deformations of the vessel wall. In order
to do this, a pressure map calculated in the previous step (fig. 3(a)) was put onto
the artery wall in the model, whereas the edges of the artery inlet and outlet were
supported (Fixed Support). Next, the calculations of the prepared strength analysis
were made. Fig. 3(b) and fig. 3(c) show the obtained map of the stress WSS reduced
according to the Huber-Mises hypothesis and a total deformation of the artery wall.
The obtained values were used in the further part of this work to conduct the anal-
ysis of the blood flow through the artery taking into account its elasticity and wall
thickness.
(a) (b) (c)
(d) (e) (f)
Fig. 29.4: Results of blood flow analysis: a) without vessel wall properties, b) with
vessel wall properties
The analysis of blood flow through the artery with a deformable wall was con-
ducted for the blood of the same properties and boundary (initial) conditions of the
flow as in the earlier analysis. Fig. 29.4b shows the results of the conducted anal-
ysis, namely the obtained wall shear stress WSS of the artery and the distribution
of pressure taking into consideration deformations of the artery wall caused by the
flow. Maximum values of the artery WSS in the case of a deformable wall are con-
siderably higher than average stresses for a rigid artery. The difference in both cases
amounts to over 30 Pa and is almost 10% higher than maximum stress WSS of a de-
formable vessel wall. The wall shear stress WSS of the vessel, without taking into
account the properties of the artery, may locally amount even to 253 Pa. This fact
may reflect the lack of adaptation (lack of deformation) of the artery to the blood
flow parameters. That is the reason why it is so important to take into account the
properties of the vessel walls in the computational fluid dynamics (CFD).
(a)
(b)
Fig. 29.5: The influence of artery wall thickness on: a) pressure, b) stress WSS
The authors also conducted simulations to evaluate the sensitivity of the sys-
tem to the wall thickness of the vessel. The response of the system (hemodynamic
parameters of vascular blood flow: pressure and stress WSS) concerning to the dif-
ferent arterial thickness was studied. According to the study [12], the average wall
thickness of the middle cerebral artery is 0,094 mm±0,025 mm. Therefore, the cal-
culations for arterial wall thickness in the range of 0,1-0,3 mm were performed in
the simulations. The hemodynamic parameters obtained for wall thickness vessel
variable are shown in fig. 29.5.
The results show an increasing of pressure along with growing of wall thickness.
However these differences amount to ±61 Pa (±2%). On the basis of the results, it
can be observed that artery wall thickness does not significantly affect hemodynamic
blood flow factors. An influence the artery wall thickness on stress WSS is similar.
29.4 Discussion and Conclusions
The results of the conducted simulations revealed the differences in the blood flow
through the artery with a rigid wall and through the vessel with a deformable wall.
The obtained hemodynamic parameters show that the vessel properties play a crit-
ical role during the blood flow in the artery. In spite of the fact that the artery is
constantly subjected to physiological pressure changes, the blood flow is stabile.
It happens due to elastic walls of the vessel which cause the decrease of pulsation
thanks to deformation. That is why it is essential that the blood flow analysis should
take into consideration the deformations of the vessel walls.
The values of hemodynamic parameters, such as: pressure, stresses WSS or ve-
locity of the flow obtained during the analysis of the artery with a non-deformable
wall are higher than in the case of elastic vessels. High values of the parameters
would suggest a risk of damage to the vessel or occurrence of an aneurysm. How-
ever, in a situation close to physiological one (a vessel with a deformable wall) such
values are considerably lower, hence there is no risk of the initiation of pathologi-
cal changes. With a view to the above, the prediction of the artery damage should
be carried out on the basis of the blood flow analysis taking into consideration the
mechanical properties of the vessel wall, which is also confirmed by other authors’
research (Alishahi et al. 2011, Sharma et al. 2004, Xiuqing Qian et al. 2011).
Acknowledgements
The study was supported by the research grant StrategMed 2/269760/1/NCBR/2015

of the National Centre for Research and Development (NCBR).
References
1. Alishahi M., Alishahi M.M., Emdad H.: Numerical simulation of blood flow in a flexible
stenosed abdominal real aorta, Scientia Iranica B, 18 (6), 1297-1305 (2011).
2. Brian D. Stemper, Yoganandan N., Stineman M. R., Gennarelli T. A., Baisden J. L. and Pin-
tar F. A.: Mechanics of Fresh, Refrigerated, and Frozen Arterial Tissue. AJournal of Surgical
Research. 139, 236-242 (2007).
3. Ferruzzi J., Vorp D. A. and Humphrey J. D.: On constitutive descriptors of the biaxial me-
chanical behavior of human abdominal aorta and aneurysms. J. R. Soc. Interface. 8 435-450
(2011).
4. Fung, Y.C.: Biomechanics. Mechanical Properties of Living Tissues, 2nd edition, Springer, New
York (1993).
5. Gzik-Zroska B., Wolański W., Gzik M.: Engineering-aided treatment of chest deformi-ties to
improve the process of breathing. International Journal for Numerical Methods in Biomedical
Engineering, Vol.29, No.9, pp. 926-937 (2013).
6. Gzik-Zroska B., Joszko K., Wolański W., Gzik M.: Development of new testing method of
mechanical properties of porcine coronary arteries, Information technologies in medicine. 5th
International conference, ITIB 2016, Kamień Ślaski,
˛ Poland, June 20-22, 2016. Proceedings.
Vol. 2. Eds. Ewa Pi˛etka, Paweł Badura, Jacek Kawa, Wojciech Wi˛ecławek. Cham : Springer,
289-297, bibliogr. 25 poz. (2016).
7. Holzapfel, G.A., Ogden, R. W.: Constitutive modelling of arteries. Proceedings of the Royal
Society A: Mathematical, Physical and Engineering Sciences, 466(2118) 1551-1597 (2010).
8. Humphrey J.D., Holzapfel G.A. Review: Mechanics, mechanobiology, and modeling of human
abdominal aorta and aneurysms. Journal of Biomechanics. 45, 805-814 (2012).
9. Lasheras, J. C.: The Biomechanics of Arterial Aneurysms. Annual Review of Fluid Mechanics.
39, 293-319 (2007).
10. Sharma G.C., Jain M., Kumar A.: Performance Modeling and Analysis of Blood Flow in
Elastic Arteries, Mathematical and Computer Modelling, 39, 1491-1499 (2004).
11. Schulze-Bauer C.A., Holzapfel G.: Determination of constitutive equations for human arteries
from clinical data. Journal of Biomechanics. 36, 165-169 (2003).
12. Walmsley J., Campling M., Chertkow H.: Interrelationships among wall structure, smooth
muscle orientation, and contraction in human major cerebral arteries. Stroke 14, 5 (1983), 781-
790.
13. Wolański W., Gzik-Zroska B., Joszko K., Gzik M., Sołtan D.: Numerical analysis of blood
flow through artery with elastic wall of vessel, [in]: Innovations in Biomedical Engineering,
Proceedings IiBE 2016, Eds. Marek Gzik, Ewaryst Tkacz, Zbigniew Paszenda, Ewa Pi˛etka,
Cham: Springer, 193-200 (2017).
14. Xiuqing Qian, Yan Wang, Zhilun Zhou, Zhicheng Liu: Effects of Material Properties on
Hemodynamic Parameters of the Coronary Artery, Computing in Cardiology; 38, 29-32 (2011).
Chapter 30
Influence of strain rates on the hyperelastic
material models parameters of pig skin tissue
Sylwia Łagan and Aneta Liber-Kneć
Abstract. To describe the mechanical behavior occurring in biological materials

during the load, the hyperelastic material models are often used. The purpose of this
work was to analyze the influence of three different strain rates on the results of
fitting the characteristic curves of selected multi-parametric hyperelastic material
models (neo-Hookean, Mooney-Rivlin, Humprey, Yeoh, Veronda-Westmann and
Ogden). Experimental data were obtained from uniaxial tensile tests of pig skin
tissue. Three values of speed were set at 1, 5, 10 mm/min. Correlation coefficients
and fitting error were evaluated. The study revealed the relationship between the
level of load speed and the values of model parameters.
Keywords: hyperelastic material models, tensile test, porcine skin
30.1 Intoduction
In order to understand the behavior of human soft tissues under different stress con-
ditions (pure shear, unconfined compression, uniaxial or biaxial tension) [2], [8],
[10], the animal substitutes are often used, among them the pigs are the most com-
monly used [4], [5], [6], [10]. The constitutive material models are used to predict
the tissue mechanical properties, which are essential to realize computer simula-
tions in surgical planning or design prosthesis of tissues. The soft tissues are often
determined as hyperelastic materials, the constitutive equations for them are usu-
ally presented by the Piola-Kirchhoff stress tensor [3] and the strain energy func-
tion. There are various kinds of the strain energy function models. Neo-Hookean,
Mooney-Rivlin, Ogden, and Yeoh are some of the most significant models, which
are the basics of the other models [4], [5], [6], [7]. To construct physical models
Cracow University of Technology, Warszawska 24, 31-155 Cracow, Poland,

e-mail: slagan@mech.pk.edu.pl,
WWW home page: http://pk.edu.pl

https://doi.org/10.1007/978-3-319-70063-2_30
280 Sylwia àagan et al.
for the identification of mechanical properties of tissue tensions, different types of

tissue were chosen to perform the mechanical tests, such as human coronary arter-
ies [3], and the human and a pig skin [4], [5], [6], [10], ligaments [1], and trachea
[9]. The aim of this work was comparison of different kinds of hyperelastic material
models to evaluate the model parameters of pig’s skin under different strain rates.
30.2 Material and Methodology
The specimens of skin tissue for examinations were taken from the back of a 9-
month domestic pig (the weight of animal was 135 kg). The surface of specimens
was cleaned from hairs, the fat was removed from the skin using a sharp scalpel
blade. The skin samples were cut in one direction, parallel in the reference to the
spine. The dimensions of the samples were: the length 100.0±0.2 mm, the measure-
ment base of samples was 50±0.2 mm, the width 10.0±0.1 mm and the thickness
was 2.0±0.1 mm. The samples were stored at the temperature of 4◦ C in polypropy-
lene closed containers, no longer than 72h before the test. The uniaxial tensile tests
were carried out with the use of the MTS Insight 50 testing machine, the range of
measuring head was 1kN, at the temperature of 22±2◦ C. The samples were mounted
using flat clamps and they were elongated at three different constant rates of 10, 5
and 1 mm/min (the strain rates of 0.0033, 0.0017, 0.0003 1/s). Each set of samples
for tension testing contained minimum 5 samples, only samples broken in the half
of measurement length were analysed. In the results, the representative stress-strain
curves were presented. Next, the neo-Hookean, Mooney-Rivlin, Ogden, Humprey,
Verona-Westmann and Yeoh models were used to describe the tensile behavior of
the pig’s skin. A hyperelastic material models are based on the definition of strain
energy function, which is expressed in different ways and depended on the class
or kind of materials considered. Assuming isotropy of material, the strain-energy
function can be written as depended on the strain invariants of the deformation ten-
sor of Cauchy-Green (I1 , I2 , I3 ). Considering the conditions of uniaxial tension of
incompressible materials the strain-energy function can be given according to the
formulas given in Table 1.
In order to obtain approximation of experimental data with mathematical record,
the fitting procedures with the use of software OriginPro7.5 were made. To deter-
mine the values of Ci algorithm of Levenberga-Marquardta (damped least-squares)
was used. For the iteration process the initial values of parameters were defined as
an equal 1. Additionally, for Mooney-Rivlin (b) model Ci ≥ 0, was defined. In this
case the model was obtained as neo-Hookean. Fitting characteristics, the coefficient
of determination R2 (1) and the model parameters Ci were recorded for two differ-
ent ranges of stretch, total 1<λ <1.7 and limited 1<λ <1.2 (the first minimum of error
–transition point).
∑nn=1 (ŷi − ȳi )2

R2 = (30.1)
∑nn=1 (yi − ȳi )2
where yi is the real value of the experimental stress, ŷi is the theoretical value of
the stress on the basis on models, ȳi is the arithmetic mean value of the experimen-
tal stress. To evaluate the match between theoretical ft (λ ) and experimental fe (λ )
stretch function at each stretch level, normalized error was calculated according to
the formula (2) [5]:
| fe (λ ) − ft (λ )|
ER(λ ) = (30.2)
fe (λ )
Table 30.1: Utilized strain energy functions formulas for an incompressible isotropic
material under uniaxial tension.
Material model Incompressible strain energy function
neo-Hookean W=C1 (I1 − 3)

Mooney-Rivlin W=C1 (I1 − 3) +C2 (I2 − 3)
Humphrey W=C1 (eC2 (I1 −3) − 1)
Veronda-Westmann W=C1 (eC2 (I1 −3) − 1) − C12C2 (I2 − 3)
Yeoh W=C1 (I1 − 3) +C2 (I1 − 3)2 +C3 (I1 − 3)3
Ogden W= α2μ2 (λ1α + λ2α + λ1−α λ2−α − 3)
30.3 Results and Discusion
All results obtained in modelling procedures, as well as the values of the coefficient
of determination R2 were shown in Table 2 and Table 3. Figures 1 and 3 showed
all stress-stretch fitting curves achieved for both ranges of stretch. It’s difficult to
compare the values of model parameters - with literature data because of different
test conditions such as strain ratio, type of samples, place and directions of speci-
mens taken, storage condition, humidity, even sex and age of animals. Error residu-
als between experimental data and theoretical results for total range of stretch were
presented in Figure 2. The initial conditions had a sufficient influence on results. For
Mooney-Rivlin model (a), for total range of stretch the values of Ci were | C1 | > |
C2 |, while for limited range of stretch and Ci ≥ 0 the values were | C1 | < | C2 |, for
all strain ratios.
For neo-Hookean, Mooney-Riviln, Humprey, and Ogden, models the values of
all model constants decreased together with decreasing the velocity of tensile tests
[6], [10]. For the Veronda-Westmann models, the C1 constant increased while the
C2 constant decreased, for the Yeoh, constant C1 and C2 decreased while the C3
constant increased and was negative. For the Ogden model the α constant changes
Table 30.2: Parameters of hyperelastic material models for skin tissue of pig for
1 < λ < 1.7.
10 mm/min 5 mm/min 1mm/min

Material model Ci R2 Ci R2 Ci R2
neo-Hookean 8.810±0.066 0.937 7.772±0.560 0.940 4.977±0.036 0.938
Mooney-Rivlin (a) 27.893±0.302 0.994 22.417±0.228 0.994 12.926±0.103 0.995
-25.452±0.402 -20.292±0.315 -12.020±0.155
Mooney-Rivlin 8.818±0.952 0.937 7.782±0.726 0.940 4.983±0.377 0.938
0 0 0
Humprey 6.072±0.285 0.980 7.172±0.333 0.981 7.023±0.307 0.981
1.049±0.036 0.788±0.027 0.503±0.016
Veronda-Westmann 18.280±0.761 0.975 21.316±0.853 0.974 21.181±0.792 0.972
0.793±0.021 0.620±0.016 0.416±0.009
Yeoh 3.332±0.074 0.997 3.032±0.068 0.996 1.814±0.043 0.996
16.617±0.289 10.997±0.201 4.285±0.078
-14.286±0.342 -7.251±0.183 -1.653±0.042
Ogden 11.743±0.185 0.983 10.528±0.164 0.983 6.606±0.138 0.984
6.894±0.083 6.350±0.071 5.696±0.054
Table 30.3: Parameters of hyperelastic material models for skin tissue of pig for
1<λ <1.2.
10 mm/min 5 mm/min 1mm/min

Material model Ci R2 Ci R2 Ci R2
neo-Hookean 5.033±0.109 0.822 3.802±0.000 0.822 1.528±0.063 0.679
Mooney-Rivlin (a) 49.445±0.200 0.999 38.519±0.156 0.999 22.773±0.348 0.988
-51.035±0.230 -39.898±0.180 -24.405±0.400
Mooney-Rivlin 5.033±3.156 0.822 3.802±2.638 0.814 1.529±1.836 0.679
0 0 0
Humprey 0.221±0.010 0.985 0.158±0.008 0.983 0.026±0.002 0.972
10.792±0.264 11.120±0.302 17.347±0.572
Veronda-Westmann 0.609±0.031 0.987 0.429±0.025 0.986 0.055±0.005 0.978
7.538±0.202 7.826±0.236 13.785±0.526
Yeoh 0.567±0.029 0.999 0.289±0.026 0.999 -0.396±0.022 0.998
45.686±0.521 36.140±0.464 16.886±0.394
-100.263±0.084 -80.860±2.507 -18.517±2.139
Ogden 3.617±0.084 0.991 2.636±0.072 0.989 0.529±0.031 0.982
18.136±0.216 18.492±0.250 25.083±0.468
Fig. 30.1: Stress-stretch curves and fitting curves for total stretch range 1<λ <1.7.
Fig. 30.2: Error residual between experimental data and theoretical results for
1<λ <1.7.
Fig. 30.3: Stress-stretch curves and tting curves for limited stretch range 1<λ <1.2.
were small enough to assume no effect of the load velocity on its value (α < μ , μ
– decreased, α – constant). For the limited range of stretch the parameters of Og-
den’s model behaved in different way (α > μ , μ – decreased, α – constant, but not
for 1 mm/min). It is with good agreement with results of Lim J., et al. [6]. Sim-
ilar values of characteristic were presented in [7] for the pig muscular tissue, for
which tensile tests were performed along the muscle fibers direction. The accuracy
of the fit also remained unchanged by the strain ratio, but varied depending on the
model. The correlation coefficient values for all models were greater than 0.9. The
highest R2 values were for Yeoh (0.997), Mooney-Rivlin (0.994), Ogden (0.9829),
Humprey (0.981), Mooney-Rivlin (0.938) consistent with neo-Hookean correlation
coefficient. Among different hyperelastic models, Yeoh and Mooney-Rivlin models
were the best able to describe the behavior of all stretch ranges and strain rates. In
both stretch ranges the number of minimum local of error residual curves were con-
sistent with numbers of parameter Ci , in all used material models that was referred
in [5], [7].
30.4 Conclusions
The researches was focused towards calibration of he selected constitutive material

models to sufficiently model behavior of skin tissue. An analysis of the stress-strain
curves under quasi-static uniaxial tensile tension realized with different strain rates
showed strong influence on nonlinear, hyperelastic nature of pig’s skin. The results
should be applied in analysis of skin tissue under different load conditions with the
use of finite element method.
References
1. Bajuri, M.N., Isaksson, H., Eliasson, P., Thompson, M.S.: A hyperelastic fibre-reinforced con-
tinuum model of healing tendons with distributed collagen fibre orientations. Biomechanics and
Modeling in Mechanobiology. 6, 15, 1457–1466 (2016)
2. Isvilanonda, V., Iaquinto, J.M., Paia, S., Mackenzie-Helnweinc, P., Ledouxa, W.R.; Hyperelas-
tic compressive mechanical properties of the subcalcaneal soft tissue: An inverse finite element
analysis. Journal of Biomechanics. 49, 7, 1186–1191 (2016)
3. Jankowska, M.A., Bartkowiak-Jowsa, M., B˛edziński, R.: Experimental and constitutive model-
ing approaches for a study of biomechanical properties of human coronary arteries. Journal of
the Mechanical Behavior of Biomedical Materials. 50, 1–12 (2015)
4. Łagan, S., Liber-Kneć, A.: Application of the Ogden model to the tensile stress–strain behav-
ior of the pig’s skin, in: Modeling and Simulations in Biomechanics. Advances in Intelligent
Systems and Computing. Springer. 526, 145–152 (2017)
5. Łagan, S., Liber-Kneć, A.: Experimental testing and constitutive modeling of the mechanical
properties of the swine skin tissue. Acta of Bioengineering and Biomechanics. in press, DOI:
10.5277/ABB-00755-2016-02 (2017)
6. Lim, J., Hong, J., Chen, W.W., Weerasooriya, T.: Mechanical response of pig skin under dy-
namic tensile loading. International Journal of Impact Engineering. 38:130–135 (2011)
7. Martins, P.A.L.S., Jorge, R.M.N., Ferreira, A.J.M.: A comparative study of several material
models for prediction of hyperelastic properties: Application to silicone-rubber and soft tissues.
Strain. 42 (3), 135–147 (2006)
8. Moerman, K.M., Simms, C.K., Nagel, T.: Control of tension–compression asymmetry in Ogden
hyperelasticity with application to soft tissue modelling. Journal of the Mechanical Behavior of
Biomedical Materials. 56, 218–228 (2016)
9. Safshekan, F., Tafazzoli-Shadpour, M., Abdouss, M., Shadmehr, M.B.: Mechanical Characteri-
zation and Constitutive Modeling of Human Trachea: Age and Gender Dependency. Materials.
doi:10.3390/ma9060456, 9, 456 (2016)
10. Shergold, O.A., Fleck, N. A., Radford, D.: The uniaxial stress versus strain response of pig
skin and silicone rubber at low and high strain rates. International Journal of Impact Engineer-
ing. 32, 1384–1402 (2006)
Chapter 31
Methodology of multicriterial optimization of
geometric features of an orthopedic implant
Małgorzata Muzalewska and Wojciech Moczulski
Abstract. The main goal of this paper is the development of a methodology of

multicriterial optimization of geometric features (the geometric form and dimen-
sion system) of an orthopedic implant responsible for supporting the reconstruction
of the anterior cruciate ligament in the knee joint. The methodology takes advantage
of the artificial immune system for optimization. To select an optimal solution two
criteria have been used: the implant should be strong enough and enable the best
blood perfusion. Two independent approaches were used for the selection: 1) based
on the calculation of the weighted sum method, and 2) based on the Pareto front.
Additionally, evaluation of the results was made using strength verification. In the
first step, the implant model was put into the bone model. In the second step of the
simulation it was pulled off the graft by the implant. Finally, results of optimisation
were assessed by medical doctors and technology engineers.
Keywords: reconstruction of the anterior cruciate ligament, orthopedic implant, Ar-

tificial Immune System, Pareto optimization, optimization of the geometric features
31.1 Origin of work
Increasing civilization progress causes an increase in the number of injuries and

injuries among people. One of increasingly frequent injuries is the break of the An-
terior Cruciate Ligament (ACL) located in the knee joint. The ligament is breaking
in most common of the sportsmen, and in amateur sports, such as skiing and soccer.
During standard ACL surgery, many complications occur [4, 13, 15].
The main goal of this paper is the development of a methodology of multicriterial
optimization of the geometric features (the geometric form and dimension system)

e-mail: {malgorzata.muzalewska,wojciech.moczulski}@polsl.pl,
WWW home page: http://www.polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_31
290 Maágorzata Muzalewska et al.
of an orthopedic implant responsible for supporting the reconstruction of the ante-

rior cruciate ligament in the knee joint.
This orthopaedic implant, whose aim is to create a “scaffold” for the grafted
tendon in a bone tunnel, which will be an implant with appropriate geometrical fea-
tures, will be made of a bioresorbable material. This idea of an innovative solution,
designed to accelerate the reconstruction of the anterior cruciate ligament, was pro-
posed by D.Sc. K. Ficek (patent PL 217 967 B1) [6] - the founder of the medical
center Galen-Ortopedia in Bieruń, Poland. The implant should not only precisely
anchor the tendon in the tunnel but, above all, - thanks to appropriately selected
geometric shape and other design features - accelerate the ingrowth of the bone
tissue into the grafted tendon. The condition for the correct ingrowth of the bone
tissue is to enable optimal perfusion of blood and other body fluids into the liga-
ment prosthesis (grafted tendon). Therefore, the geometrical shape of the implant
should allow accumulating active biological factors in the vicinity of the grafted
tendon, thus accelerating the healing process. In addition, the implant is to be a kind
of a transitional layer between the tendon and the surface of the bone tunnel with an
additional task to protect the bone opening against mechanical degradation caused
by the movements of the tendon (longitudinal movements when moving the knee).
Therefore, there is a need to reach a compromise between the following conditions:
providing the optimal blood flow through the implant, by simultaneously preserving
the strength properties.
The problem dealt with in the paper belongs to classical shape optimization prob-
lems that are extensively researched in many works [5]. However, in these reports
usually single-criterial optimisation problem is solved. Furthermore, the authors did
not find any report of research that concerned optimization of orthopaedical im-
plants. Thus the approach presented in the paper seems to be original.
31.1.1 Multicriterial optimization
Since a satisfactory mathematical model, which comprises both the problems of

proper blood flow through the implant and the implant strength in case of overloads
in a knee joint has not been developed yet, it is planned that Artificial Immune Sys-
tems from heuristic algorithms class will be used for the multicriterial optimization.
The goal of the multicriterial optimization is to find an optimal solution due to some
stated criteria. Choosing the optimal solution can be done in different ways, e.g.:
• by means of a weighted sum method [8],

• by using a concept of Pareto optimality [11].
In this work both the approaches were used and their results compared.
31 Methodology of multicriterial optimization 291
31.1.2 Immune algorithms
An idea of an Articial Immune System is based on the biological immune system.

The function of the biological immune system is to protect the body, and to de-
velop antibodies to ght foreign organisms (antigens). The Articial Immune Sys-
tem takes advantage of such functionalities as: innate immunity, adaptive immunity,
clonal selection, negative selection and immune network algorithms [1, 14, 2, 7].
The most important properties of the Articial Immune System are its memory and
learning ability. In general, the algorithm of Articial Immune System consists of
the following steps [10]:
• Generating a new population P of individuals;
• Selecting n individuals with the highest degree of tness;
• Cloning each of the selected individuals in proportion to the degree of tness;
• Performing hyper-mutation of cloned individuals with the probability inversely
proportional to the degree of tness;
• Second selection of improved individuals, setting up one set of memory cells
M;
• Replacement of a specied number d of individuals of a population P by better
clones - individuals with a lower degree of tness are more likely to be replaced.
31.2 Methodology of the optimal selection of geometric features

of an orthopedic implant
The methodology determines the way of proceeding used for the purpose of de-
termination of optimal geometrical features of an orthopedic implant used for the
reconstruction of the anterior cruciate ligament.
31.2.1 Design assumptions
The design assumptions are:

• Implant made of bioresorbable material - polylactide (LATI Latigea B01 F1
Bio-Polymer) [9],
• Specic dimensions of the implant,
• Application of heuristic algorithms,
• Optimization due to two criteria: strength and blood ow.
31.2.2 Criteria of concept evaluation - Strength
To keep stresses of the material in optimal interval, it was decided to depend (31.1):
Permissible stresses
f1 = . (31.1)
Calculated stresses
where permissible stresses its 5,5 MPa.
31.2.3 Criteria of concept evaluation - blood flow
In order to obtain the optimal soaking implant of the blood, it was decided to deter-
mine the hypothetical value of the desired fluid velocity at the output of the implant
(2 m3 /s) and apply dependencies (31.2) or (31.3).
If calculated speed value is greater than the predetermined speed of the blood, then:
2
Predetermined speed
f2 = . (31.2)
Calculated speed
Otherwise:
2
Calculated speed
f2 = . (31.3)
Predetermined speed
This relationship allows to find the implant characterized by a flow rate close to a
predetermined average speed of blood flow. The multicriteria optimization method
developed in this study is universal, so if the optimal blood speed would be deter-
mined in the future, the algorithm did not require any change apart of substituting
the appropriate value in the formulae. The input flow rate at the input is 1 mm/s.
31.2.4 Objective function
In the first version of optimization, the weighted-sum objective function defined by

the general formula (31.4) was used:
F(x) = w1 ( f1 (x)) + w2 ( f2 (x)), (31.4)
where: x is a vector of solutions, the index 1 defines parameters related to the stress
properties (31.1), and the index 2 defines parameters related to blood perfusion by
the implant (cf. formulae 31.2 or 31.3).
In this paper, the objective function is defined as follows:
If calculated speed value is greater than the predetermined speed of the blood,
then:

Permissible stresses 2 Predetermined speed 2
f = w1 · + w2 · , (31.5)
Calculated stresses Calculated speed
Otherwise:
2 2
Permissible stresses Calculated speed
f = w1 · + w2 · , (31.6)
Calculated stresses Predetermined speed
31.2.5 Pareto Method
Multicriterial optimization using the Pareto method allows to find a set of possible
solutions that meet all the given criteria [11].
For each individual, the factor of stress is used according to the formula (31.1)
and the blood perfusion formula (31.2) or (31.3). These values are then checked
according to (31.7):
F(x) = ( f1 (x), f2 (x), ..., fn (x)). (31.7)
31.3 Multicriterial optimization method using Immune

Algorithm
To verify the methodology a software environment was built using optimization

method based on the artificial immune system. This software was developed in
MATLAB. MATLAB software works on batch files - dedicated to ANSYS APDL
(strength calculations) and ANSYS CFX (blood perfusion calculation).
The steps of the implemented immune algorithm are:
1. Generate initial population;
2. Update the parameters in scripts dedicated to ANSYS APDL;
3. Start the batch files for calculations:
• APDL script to calculate the mechanical properties;
• APDL Script to generate a model for the calculation of the perfusion;
• CFX-Pre script to define the initial parameters required for the calculation
of perfusion;
• Script to launch the calculation of the - CFX-solution flow;
• The CFX-postscript module data analysis;
4. Import data from calculations of static structural analysis and blood perfusion;
5. Calculate the fitness function or Pareto-optimal indywiduals;
6. Select the best individuals (geometric form of implants);

7. Stop condition or generate the next population and go to the 2nd step.
31.4 Verification and validation of the described methodology
In order to verify the methodology the software environment described above has
been extensively applied. Additionally, evaluation of the results was made using
strength verification. To this end, in the first step, the implant model was put into
the bone model. In the second step of the simulation it was pulled off the graft by
the implant. Finally, the results were assessed by medical doctors, and technology
experts. For this purpose special forms were applied.
31.4.1 Optimisation results by using the objective function
The algorithm made it possible to find an optimal result after 21 major iterations
and completed the calculation because no significant improvement in the results
in subsequent iterations was achieved. The graph below (Fig. 31.1) illustrates the
process of optimization in subsequent iterations.
Fig. 31.1: The optimization process using the objective function and optimal geo-
metric features of the implant selected by the immune algorithm using the objective
function
The red line shows the mean of the goal function calculated for all individuals in
the given iteration, while the blue one shows the value of the objective function
calculated for the best individual in each iteration. The geometry of the optimal
implant is shown in Fig. 31.1. This implant is characterized by densely distributed
openings, in rows offset relative to each other.
31.4.2 Simulation results using the Pareto front
The Pareto function returned 149 results, including 8 non-dominated constituents

forming the Pareto front. In the figure 31.2 the dominated individuals are marked
in blue, while the Pareto front is marked in red. In the picture 3 shows the implants
Fig. 31.2: Geometric features of Pareto-optimal individuals
diffrent the number of side openings, the length, width and the relative position to
each other.
31.4.3 Strength verification
The best solution found by means of the multicriterial optimization has been then
verified against loads that are created after the orthopaedic operation, incase of walk-
ing. Here were modeled bones of the knee joint. In this bone, basing on DICOM
images, the cortical and soft bone tissues were separated. The results of the strength
verification showed that most of the loads carries the hard tissue, so the loads go-
ing to the soft tissue and the implant are small. Mechanical properties of bones are
shown in tab. 31.1 [12]. The Pauwels model was used to determine the load of the
femur. Fixation, load of bone and stress distribution in the bone is shown in figure
31.3.
Table 31.1: Mechanical properties of femur [12]
Mechanical properties Cortical bone Soft bone

Young’s module [MPa] 17 000 300
Density [g/cm3 ] 1,9 0,46
Poisson’s ratio 0,3 0,3
Compressive strength [MPa] 200 6
Fig. 31.3: Stress in bones
31.5 Conclusion
The main goal of the work described in the paper was the development of a method-
ology of multicriterial optimization of geometric features (geometric form and di-
mension system) of an orthopedic implant responsible for supporting reconstruction
of the anterior cruciate ligament in the knee joint. The main part of the work is the
description concerning the methodology of optimization of the geometric features
of an orthopedic implant, and selection of the optimal solution with respect to the
following two criteria: the form geometry should be strong enough and enable the
best blood perfusion. Two independent approaches were used for the selection, the
first being based on the calculation of the weighted sum method, the second based
on the Pareto front. Both the methods gave similar results. Which means, that the
geometric figures of implants are similar.
The analysis of the results confirmed the thesis: multicriterial optimization sup-
ported by: the artificial immune system, the weighted sum method and Pareto op-
timization, make it possible to choose the geometric features which allow optimal
blood perfusion while providing the optimum strength of the implant.
References
1. Al-Enezi J.R., Abbod M.F., Alsharhan S. Articial immune systems - models, algorithms and
applications. IJRRAS 3 (2), 118-131. May 2010.
2. de Castro L.N., Von Zuben, F.J. Immune and Neural Network Models: Theoretical and Em-
pirical Comparisons. International Journal on Computational Intelligence and Applications,
2001, vol. 1. no. 3. p. 239–257.
3. Dáugosz A. Optymalizacja wielokryterialna w problemach pól sprzez onych. Wydawnictwo
Politechniki Ślaskiej,
2013.
4. Domnick Ch., Raschke M. J. and Herbort M. Biomechanics of the anterior cruciate ligament:
Physiology, rupture and reconstruction techniques. World Journal Orthopedics, 2016, vol. 7.,
p. 82–93.

5. FedeliÅDski P., GÃşrski R. Optimal arrangement of reinforcement in composites. Arch. Civ.
Mech. Eng. 2015 vol. 15 iss. 2, p. 525-531.
6. Ficek K. Patent nr PL217967 B1: Implant medyczny do wzmacniania wgajania przeszczepów
w rekonstrukcji wiezadeá w tunelach kostnych. 30.09.2014.
7. Gong M., Liu C., Jiao L., Cheng G. Hybrid immune algorithm with Lamarckian local search
for multi-objective optimization. Memetic Computing, 2010, vol. 2., p. 47–67.
8. Kaddani S., Vanderpooten D., Vanpeperstraete JM., Aissi H. Weighted sum model with par-
tial preference information: Application to multi-objective optimization . European Journal of
Operational Research, 2017, vol. 260, iss. 2, p. 665-679.
9. Matweb material property data. 2016. http://matweb.com/search/DataSheet.
aspx?MatGUID=4a4781f0791445a5aff417ed6f778737&ckck=1.
10. Świtalska A. Sztuczne systemy immunologiczne - zastosowanie w optymalizacji kombinato-
rycznej. Instytut Podstaw Informatyki Polskiej Akademii Nauk, 2006.
11. Tarnowski W. Optymalizacja i polioptymalizacja w technice. Wydawnictwo Uczelniane Po-
litechniki Koszalińskiej, 2011.
12. Tejszerska D., Świtoński E., Gzik M. Biomechanika narzadu ruchu czáowieka. Instytut Tech-
nologii Eksploatacji - PIB, Radom, 2011.
13. Weiler A., Peine R., Pashmineh-Azar A., Abel C., SÃijdkamp N. P., Hoffmann R. F., Tendon
healing in a bone tunnel. Part I: Biomechanical results after biodegradable interference t
xation in a model of anterior cruciate ligament reconstruction in sheep. Arthroscopy, 2002,
vol. 18(2), p. 113–123.
14. Wierzchoń S. T. Sztuczne systemy immunologiczne. Teoria i zastosowania. EXIT, Warszawa
2001.
15. Wing-Yuik I., and Gogolewski S. Clinical application of resorbable polymers in guided bone
regeneration. Macromolecular Symposia, 2007, vol. 253, p. 139–146.
Chapter 32
Research on the stability of the users of chair
with a spherical base
Robert Michnik1 , Miáosz Chrzan1 , Piotr Wodarski1 , Andrzej Bieniek1 , Katarzyna

Nowakowska1 , Anita Pollak2 , Andrzej Mitas3
Abstract. One of the way to get rid of the spine issues is using chairs with a movable
seat. The essence of chairs with a movable seat is to increase the range of motion
of pelvis movement. Natural ability to maintain the upright position of the body can
cause the increasing of pelvis movement. That requires the involvement of the spine
stabilizing muscles. The construction of chairs with a spherical base should provide
stability and prevent from slipping away from the chair. At the same time chairs
with spherical base should provide mobility to force pelvis movements. We made
an ergonomic analysis of chair with spherical base. Studies have determined the
stability of the seat and the range of pelvis movements. These measurements used
the whole body kinematics. The ergonomic assessment of the spherical seat required
investigation to determine the following parameters: center mass displacement, lean
of spherical chair and kinematics of the human body.
Keywords: active chair, center mass displacement, sitting, stability
32.1 Introduction
Statistical data indicate that ailments and disorders within the spinal area con-
stitute the most common group of chronic conditions. Nowadays, spinal pains are
often considered to be a civilization disease. There has been a considerable increase
in the number of people affected by body posture defects in recent years [3]. There
are many factors contributing to the occurrence of pain and ailments: a sedentary
lifestyle, lack of physical activity, obesity, bad habits or a body posture that was not
Department of Biomechatronics, Faculty of Biomedical Engeenering, Silesian Univesity of

Technology, Poland
e-mail: Robert.Michnik@polsl.pl,
WWW home page: https://www.ib.polsl.pl · Institute of Psychology, University of Sile-
sia, Katowice, Poland · Department of Informatics and Medical Equipment, Faculty of Biomedical
Engeenering, Silesian Univesity of Technology, Poland

https://doi.org/10.1007/978-3-319-70063-2_32
300 Robert Michnik et al.
corrected at a young age. Thanks to rapid development of medical sciences in recent

years, many new methods of treatment and rehabilitation of spinal conditions have
come into being. However, all activities that aim to avert ailments also play a sig-
nicant role in pain prevention. All preventive measures should be rst of all based
on the care of the spine work ergonomics in daily life, i.e. they should include the
formation of a correct body posture as well as the strengthening of the muscles sta-
bilizing the spine. One of the examples of preventive activities is the use of special
chairs with a movable seat.
Sitting is one of the daily positions which are the most straining for the spine.
During the action of sitting and getting up from the chair the values of a resultant
force of the reaction in the lumbar spine section may constitute even 180% of the
body weight [4], [5]. The essential feature of the chair with a movable seat is the fact
that it increases the range of the pelvis movement or forces the pelvis movement.
Natural ability or pursuit for maintaining an upright position of the body result in the
fact that the increase of the pelvis mobility requires the engagement of the muscles
supporting the spine. Greater mobilization of stabilizing muscles contributes to their
strengthening and this in turn guarantees the improvement of keeping the balance
and prevents the defects of the body posture [10], [1], [7]. Today, there are two types
of chairs with a movable seat available on the market, i.e.: chairs with a seat mounted
on a spring and chairs with a spherical base. Increased mobility of the seat has a lot
of pro-health advantages in comparison with standard chairs, however, it may also
pose a greater risk of losing stability or falling off the chair. While using the chair
with the seat based on a spring, the fall may occur as a result of sliding from the
seat at a too high angle of inclination from the horizontal plane. In the case of the
chairs with a spherical base, the loss of stability will result from the displacement
of the centre of mass beyond the plane of support. It may be thus concluded that
the chairs with a spherical base pose a greater risk of fall.The construction of chairs
with a spherical base should provide proper stability and prevent the probability
of sliding from the chair. At the same time, it should ensure adequate mobility of
the seat in order to force movement of the pelvis and activate stabilizing muscles.
Biomechanical tests of the above-mentioned seats were performed by Burg J.C.E.
et al. as well as Willigenburg M.W. et al. [10], [11]. The main purpose of their
research was the analysis of the displacement of the centre of pressure. There are
yet very few papers that focus on the analysis of motion of the whole human body
during the use of unstable seats. Accessibility of the Motion Capture systems makes
it possible to evaluate the movement of the whole human body as well as the impact
of its individual segments on keeping balance. Possibility of the determination of
the motion of the pelvis and individual segments of the spine allows the verication
of the correctness of interaction and impact of movable seats on the human motor
system [9].
This work aimed to conduct research on the kinematics of the users of the chair
with a spherical base as well as to determine its stability and the range of the pelvis
movement while sitting in such a chair.
32 Stability of chair with a spherical base 301
32.2 Research methodology
The research on the kinematics of the users of chairs with a spherical base
used an inertial system for motion analysis the MVN Biomech of the XSENS com-
pany. The system consists of 17 inertial sensors which are placed in certain spots
on the body of the examined person and are linked to two transmitters which wire-
lessly send a signal to receivers. Each transmitter is equipped with an accelerometer,
gyroscope and magnetometer. The scope of operation of the MVN BioMech gyro-
scopes amounts to ±1200◦ /s, whereas in the case of accelerometers to ±16g. The
frequency of the data recording was 120 Hz. The MVN Biomech measuring sys-
tem enables the determination of linear displacement of individual anatomic points
(such as joints and mass centres), angles in joints, angles dening the orientation of
the body segments in space as well as derivatives of the above-mentioned values.
The test group consisted of eight healthy men from 21 to 24 years of age. The
methodology of experimental testing encompassed the performance of a dozen or so
measurements for each examined person. During the tests the participants conducted
different activities while sitting in a chair with a spherical base. The tests measured
the kinematics of the body motion during the following activities:
- leaning out of the chair in the transverse and sagittal axes of the human body
within the limits of comfort (testing the limits of comfort),
- leaning out of the chair in the transverse and sagittal axes of the human body in
order to lose balance (testing the limits of stability),
- free sitting,
- sitting while having a conversation.
Figure 1 presents a person examined during experimental measurements con-
ducted with the MVN Biomech uniform for motion analysis, whereas gure 2 shows
an avatar of a person examined on the basis of the MVN Studio software during a
maximum deection in the transverse and sagittal axes of the body.
The following values were subjected to a detailed analysis aiming at the assess-
ment of ergonomics and stability of the tested seat: the inclination angle of the chair
with a spherical seat in relation to transverse plane, the inclination angle of the pelvis
in relation to transverse plane and the displacement of the mass centre of the human
body.
32.3 Results
The results of the displacement of the body mass centre (mean ± standard de-
ection) obtained while testing the limits of stability and comfort as well as free
sitting and sitting during a conversation were presented in Table 1. Figure 3 shows
examples of the displacement courses of the body mass centre for one of the partic-
ipants during the tests of stability limits.
The average displacement of the body mass centre of the tested persons for all
analyzed activities was presented in Figure 4. The points of maximum mean dis-
Fig. 32.1: Experimental measurements conducted with the MVN Biomech uniform
for motion analysis
Fig. 32.2: Mesurements of the leaning out of the chair in the transverse and sagittal
axes of the human body within the limits of comfort
Table 32.1: Displacement of the body mass centre obtained while testing the limits
of stability, limits of comfort, free sitting and sitting during a conversation
Right - Left - Forward - Bacward -

displacement [m] displacement [m] displacement [m] displacement [m]
Limits of stability 0.147 ± 0.034 0.134 ± 0.060 0.237 ± 0.039 0.198 ± 0.044
Limits of comfort 0.087 ± 0.015 0.089 ± 0.026 0.132 ± 0.040 0.096 ± 0.035
Free sitting 0.024 ± 0.028 0.024 ± 0.028 0.015 ± 0.014 0.016 ± 0.012
Conversation 0.021 ± 0.022 0.012 ± 0.006 0.009 ± 0.006 0.013 ± 0.010
Fig. 32.3: Displacement of the body mass centre obtained while testing the limits of
stability
placements were linked to each other by means of lines and thus a support rhombus
was created.
Fig. 32.4: The average displacement of the body mass centre of the tested persons
For the conducted tests of stability and comfort limits as well as free sitting and
sitting while conversing the following parameters were also subject to analysis: the
inclination angle of the pelvis and the inclination angle of the seat. The obtained
results were tabulated in Table 2.
Table 32.2: Range of the inclination angle of the pelvis and the inclination angle of
the seat
Left-Right Forward-Backward
range of lean [degree] range of lean [degree]
Limits of stability 32.1 ± 8.8 36.3 ± 12.6
Pelvis
Limits of comfort 30.2 ± 9.4 34.2 ± 6.3
Limits of stability 19.3 ± 8.4 23.6 ± 10.5
Seat
Limits of comfort 12.8 ± 6.9 13.0 ± 9.0
Free sitting 1.5 ± 1.2 1.5 ± 1.7
Pelvis
Conversation 0.6 ± 0.7 0.3 ± 0.2
Free sitting 0.6 ± 0.7 0.5 ± 0.4
Seat
Conversation 2.0 ± 1.3 1.2 ± 0.9
32.4 Discussion
A sedentary lifestyle, a rapid increase in the occurrence of spinal ailments

as well as a greater awareness in the scope of health care proved the need for the
replacement of regular traditional chairs with pro-health alternative chairs, the so-
called active chairs. There are many solutions and models of active chairs on the
market, however, there is a lack of objective biomechanical evaluation of the opera-
tion of particular types of seats.
Research papers concerned with the testing of sitting on movable seats mainly
focus on the analysis of the displacement of the centre of pressure. This parameter
was determined for the group of healthy persons among others by van der Burg
[10] in the research on the sitting on an active seat with a spherical base. In Burg’s
investigation a maximum displacement of the pressure centre along the transverse
axis of the body amounted to 0.016 m, whereas in the presented study a maximum
displacement of the body mass centre while sitting on the analyzed seat was equal
to 0.024 m. For the motion in the sagittal axis of the body, a maximum value of
the displacement of the body mass centre amounted to 0.016 m in the presented
study, whereas in Burga’s investigation the displacement of the centre of pressure
equalled 0.020 m. The differences in the obtained results were probably the reason
for a different diameter of the sphere used for making the seat. They may also result
from the difference of distance between the spherical cap and the position of the
pelvis.
The values of displacement of the body mass centre, which were obtained in the
presented study, were compared to the results obtained during stabilographic tests.
An average range of deflections in the sagittal plane while standing in the research
conducted by Piecha M. et al. equalled 0.023 ± 0.007 m. For the frontal plane this
scope was equal to 0.019 ± 0.007 m [6]. Piecha M. et al. obtained the values of
the displacement of the body mass centre in the sagittal axis and frontal axis which
were almost by 50% smaller than the values obtained in the authors’ own research
for free sitting on the chair with a spherical base.
While analyzing the values of the displacement of the body mass centre in the
tests of stability limits, it was observed that smaller displacements occurred for de-
flections of the human body in the transverse axis (deflections in the right and left di-
rections). Deflections in the sagittal axis of the human body were on average higher
by 0.077 m. A difference was noted also between deflections in the sagittal axis of
the human body. During deflection in the forward direction, an average value of the
displacement of the mass centre was by 20% higher than in the case of backward
deflection. It resulted most probably from the construction of the chair or the selec-
tion of a site where a given person was sitting in the chair. The displacement of the
body mass centre during deflections in the transverse axis was almost symmetrical.
Moreover, it should be emphasized that during losing balance the position of lower
limbs prevented falling in the transverse axis (Fig. 2). Further parameters which
were analyzed in the testing of stability limits were the angle of the pelvis inclina-
tion as well as the angle of the seat inclination. It was observed that the inclination
of the pelvis and the seat was higher during deflections in the sagittal axis than in the
transverse axis. For forward and backward deflections the values of the angle of the
pelvis inclination were higher by 13%, whereas for the angle of the seat inclination
by 22% in relation to side deflections.
The analysis of the results obtained in the testing of comfort limits revealed that
the displacement of the body mass centre was by 30% higher for deflections in
the sagittal axis than in the transverse axis. In addition to that, forward deflections
were higher by 37.5% in relation to backward deflections. Similarly to the tests of
stability limits, the displacements of the body mass centre in the transverse axis
were symmetrical. The angle of the pelvis inclination was higher by 13% during
deflections in the sagittal axis than in the transverse axis. No differences were found
for the values of the angle of the seat inclination while testing comfort limits.
While comparing the obtained values of the motion ranges of the body mass
centre in the tests of stability and comfort limits, it was observed that the results
acquired for comfort limits were almost twice as lower in relation to the results
obtained during the testing of stability limits. It reflects an intuitive safe use of the
seat without losing stability by participants from the test group.
32.5 Conclusions
Testing of the safe use of the chair should constitute an important aspect of the
designing process of chairs with a movable seat. The testing methodology presented
in this work makes it possible to determine the limits of comfort and stability while
sitting in the chairs with a spherical base.
The obtained results indicate that intuitive sitting in the chair subject to analysis
is safe for the users. The values of the displacement of the body mass centre obtained
while testing comfort limits are almost twice as lower as the results obtained while
testing stability limits. On the basis of the acquired results of the displacement of
the body mass centre as well as the results obtained by other researchers, it was
indicated that the displacements of the centre of mass while sitting on movable
seats are very little. It seems justified that a similar methodology, which was used
in stabilographic tests, should be applied to the analysis of the displacements of the
body mass centre. Such methodology includes the determination of the path length,
the surface area of the ellipse, the velocity of displacement of the centre of mass or
the application of the analysis of frequency of the obtained signals [2].
The results acquired from experimental testing make it possible to declare that the
scope of motion of the seat in the chair subject to this analysis during free sitting and
sitting while conversing is very small. A small range of the seat’s motion translates
into a small range of movement of the pelvis and the spine, and thus the effect of
strengthening of the stabilizing spinal muscles may be very little. This conclusion
has been partly confirmed by the results obtained by Szurmik T. et al. [8]. They
observed that a several-month use of a chair of this type had little contributed to the
improvement of the ability of balance keeping [8]. It has also been observed that the
construction of the seat subject to testing as well as the location of the top of the
spherical base in the rear part of the seat increases the risk of losing stability while
making a backward deflection of the body.
The obtained results constitute essential guidelines which should be applied in
the process of designing of the seats with a spherical base.
References
1. Cholewicki J., Polzhofer G.K., Radebold A.: Postural control of trunk during unstable sitting.
J Biomech, 33, 1733–1737 (2000)
2. Michnik R., Jurkoć J., Wodarski P., Gzik M., Jochymczyk-Woźniak K., Bieniek A.: The influ-
ence of frequency of visual disorders on stabilographic parameters. Acta of Bioengineering and
Biomechanics, 18(1),25–33 (2016)
3. Myśliwiec A., Jurkojć J., Michnik R., Nowakowska K., Rybok K., Posl˛uszny A.: Differences in
lumbar pain intensity between parents of children with motor disability and parents of children
without disabilities. Aktualne Problemy Biomechaniki, 10, 37–42, 2016 (in Polish)
4. Nowakowska K., Gzik M., Michnik R., Myśliwiec A., Jurkojć J., Suchoń S., Burkacki M.: The
loads acting on lumbar spine during sitting down and standing up. Innovations in biomedical
engineering, Springer International Publishing, 169–176 (2017)
5. Nowakowska K., Michnik R., Myśliwiec A., Chrzan M.: Impact of strengthening of the erector
spinae muscle on the values of loads of the muskuloskeletal system in the lumbar spine section.
Engineering Mechanics 2017, 23rd International Conference May 15 - 18, 2017, Svratka, Czech
Republic, Book of full texts, editor: Vladimir Fuis, 718–721 (2017)
6. Piecha M., Król P., Juras G., Sobota G., Polak G., Bacik G.: The effect of short- and long-term
vibration trainingon postural stability in men. Acta of Bioengineering and Biomechanics, 15,
29–36 (2013)
7. Silfies S.P., Cholewicki J., Radebold A.: The effects of visual input on postural control of the
lumbar spine in unstable sitting. Hum Mov Sci, 22, 237–252 (2003)
8. Szurmik T., Kurzeja P., Bibrowicz K., Hadlich R.: Valuation of the impact of balance exercises
with the "kivak" furniture set on the improvement of the selected stabilometric factors in 7-
year-old children. Journal of Education, Health and Sport, 12, 424–440 (2016)
9. Takei M., Shimizu H, Hoshiyama M.: Kinematicanalysis of seatingmaneuver: Digitalization of
movement in dailyliving. Human Movement, 11, 114–118 (2010)
10. van der Burg J.C.E., van Wegen E.E.H, Rietberg M.B., Kwakkel G., van Dieen J.H.: Postural
control of the trunk during unstable sitting in Parkinson’s disease. Parkinsonism and Related
Disorders, 12, 492–498 (2006)
11. Willigenburg N.W., Kingma I., van Dieen J.H.: Center of pressure trajectories, trunk kine-
matics and trunk muscle activation during unstable sitting in low back pain patients. Gait &
Posture, 625–630 (2013)
Part IV
Engineering of biomaterials
Chapter 33
Assessment Of Hip Endoprosthesis Cups After
Reimplantation
Anita Kajzer1 , Ewelina Głab

˛ 1 , Wojciech Kajzer1 , Tomasz Wróbel2 , and Jacek
Semenowicz 3
Abstract. The paper presents results of the assessment of the absence of tissue
hypertrophy to the surface of the hip endoprosthesis cups. The cups were obtained
as a result of revision operations, which caused their loosening. The study included
analysis of chemical composition, SEM, macroscopic evaluation of surface, pitting
corrosion resistance and hardness. Based on the result obtained, it can be concluded
that the surface condition of the outer part of the cups affects bone tissue hypertro-
phy to the ground.
Keywords: corrosion resistance, SEM, hip endoprosthesis cups
33.1 Introduction
Different factors lead to change in the area of the hip joint, whose consequence is
primary or secondary osteoarthritis [1]. The most common causes secondary os-
teoarthritis are hip dysplasia, avascular necrosis of femoral head, inflammatory dis-
eases, most often rheumatoid arthritis, completed injuries in the area of acetabulum
of the hip joint or in the proximal extremity femur, mainly resulted fracture femoral
neck [2, 3, 4]. Golden standard in orthopedics, in treatment osteoarthritis is his
arthroplasty [3, 5]. Early and distant results of hip arthroplasty are very good elim-
inating the pain, improving mobility and locomotive possibilities of limbs. How-
ever, complications are possible early and distant. The most common complication
is aseptic loosening the parts artificial joint, more often it’s acetabular component.
Treatment of the complication is exchange procedure of loosening elements [1, 6].
Department of Biomaterials and Medical Devices Engineering, Faculty of Biomedical Engineer-

ing, Silesian University of Technology, ul. Roosevelta 40, 44-800 Zabrze, Poland,
e-mail: anita.kajzer@polsl.pl · Department of Foundry, Faculty of Mechanical Engineer-
ing, Silesian University of Technology, Towarowa 7, 44-100 Gliwice, Poland. · Dr Janusz Daab
Independent Public Regional Hospital of Trauma Surgery in Piekary Slaskie, 62 Bytomska Street
41 - 940 Piekary Ślaskie,
˛ Polska

https://doi.org/10.1007/978-3-319-70063-2_33
312 Anita Kajzer et al.
Implants implanted into the human body are exposed to a variety of conditions in
the body [7, 8, 9, 10, 11]. Staying in the environment of tissues and body fluids, they
must be characterised by appropriately selected chemical composition, mechanical
properties and surface condition [12, 13, 14, 15, 16]. Due to the adaptation of the
implant to the damaged environment, the process of osseointegration of bone tis-
sue and implant surface is important. The growth of bone tissue depends largely on
the size of the pores of the implant surface. Their optimal "architecture" should, as
closely as possible, follow the Haversian canals [17]. At pore diameters of less than
95 Âţm, fibrous tissue is initially formed. Optimal growth takes place at sizes from
100 to 150 μ m and fastest at up to 1000 Âţm [18]. It is important to take into account
the size of the connections between the pores. Currently, the most preferred solution
is material with a pore size of 100-500 μ m and a connection between the pores of
above 50 μ m [17, 18]. In order to eliminate errors arising at the stage of design,
fabrication and implant usage, an appropriate assessment of the product should be
carried out. Therefore, the aim was to assess the cause of lack of bone biopsy of two
types of metal hip endoprosthesis cups after reimplantation.
In order to conduct the tests, four hemispherical hip endoprosthesis cups, were used.
The first group consisted of uncemented, less porous cups, with diameter of 50 mm,
fixed using press-fit technique, made of Cr-Ni-Mo steel with increased nitrogen con-
tent, with predicted chemical composition and mechanical properties in accordance
with ISO 5832 - 9. These implants are labelled as 1 and 2 (Fig.33.1ab). The second
group consisted of uncemented cup, fixed using press-fit technique with additional
screw fixation and uncemented cup fixed using press-fit technique with additional
cement bone between it and the metal superstructure stabilized with the pelvic bone
using screws, both implants with diameter of 62 mm. These implants were made of
Ti6Al4V titanium alloy, with predicted chemical composition and mechanical prop-
erties in accordance with ISO 5832 - 3, and are labelled as 3 and 4 (Fig.33.1cd). The
group of recipients of implants consisted of women. Occurred loosening of the hip
endoprosthesis cups: implant 1 - revision with replacement of the cup 109 months
after primary arthroplasty, implant 2 - arthroplasty revision in 132 months after pri-
mary surgery, implant 3 - cup removed in 7 months after the third revision surgery
and implant 4 - cup with superstructure removed in 7 months after the third revision
surgery. According to the manufacturer’s data, a 0.35 mm thick, porous titanium
layer was produced on the pan no. 3, with a pore size ranging from 50μ m to 200μ m
[19].
Macroscopic evaluation of the surface was performed using a stereoscopic mi-

croscope - SteREO Discovery.V8 from Zeiss with AxioVision software, with total
magnifications: 1.6x, 4x, 4.8x, 6x, 6.3x and 10x. The observations were made before
and after the pitting corrosion resistance test.
33 Efect Of Sterylization 313
Fig. 33.1: Material for research - hip endoprosthesis cups: a) 1, b) 2, c) 3, d) 4
In order to confirm the compliance of chemical composition with standards, spec-

troscopy and SEM using the EDS detector were carried out. Chemical composition
test for cups 1 and 2 was performed using the Leco GDS500A emission spectrome-
ter. The device based on the technique of optical emission spectrometry with fluores-
cence excitation. The following spectrometer parameters were used: voltage 1250
V, current 45 mA, argon pressure 2 Tr and vacuum 0.1 Tr. On the other hand, for
cups 3 and 4 approximate chemical composition was evaluated using the Phenom
World’s Phenom proX scanning electron microscope, integrated with the EDS. The
cups’ surface was analysed for pore size. The zoom range of the device ranged from
20x to 130000x.
The potentiodynamuc test was performed as recommended by ISO 10933-15
standard on the test stand which consisted of: the Radiometer’s VoltaLab PGP 201
potentiometer, an electrochemical cell, active electrode (of the test sample), auxil-
iary electrode (platinum electrode), reference electrode (saturated calomel electrode
SCE) and a computer with VoltaMaster 4 software. The test was started with deter-
mining open circuit potential (EOCP ) in the non-current conditions. Subsequently,
the polarisation curves were recorded, starting from the initial potential value Estart
= EOCP - 100 mV. The change of the potential occurred in the anode direction at the
rate of 3 mV/s. Once the anode current density reached j=1mA/cm2 the polarisa-
tion was changed to cathode direction. The test was performed in Ringer’s solution
simulating the environment of tissues and body fluids. Measurements were made in
250 ml of liquid at the temperature T = 37±1◦ and the pH = 7.0±0.2◦ . On the basis
of the recorded curves the breakdown potential Eb mV, the repassivation potential
Ecp mV, the transpassivation potential Etr mV and the corrosion potential Ecorr mV
were determined along with the value of the polarisation resistance R p kΩ · cm2 ,
calculated with the use of Stern method.
Hardness was measured using the Vickers method as recommended by PN-EN
ISO 6507-1:2005 standard. The measurements were carried out with the use of the
Struers’ DuraScan 50 hardness meter, with a load of 0.1 HV - for samples from the
cup number 4 because of their largest porosity compared to the other cups, and 0.5
HV for the other samples. Tested were four samples in the form of lateral transplants
of the central parts of the cups (Fig.33.1a).
The surfaces of Cr-Ni-Mo steel implants did not show any areas where the bone
tissue hyperplasia occurred. However, uneven local residues of unresorbed hydrox-
yapatite were noted (Fig.33.2ab). For titanium alloy cups, a greater porosity of the
surface was observed. On the outer surface of cup 3, the remains of connective
tissue and blood clot were observed (Fig.33.2c). In the case of implant 4, locally
occurring areas were observed, in which bone trabeculae, resulting from the onset
of bone tissue surface overgrowth, were present (Fig.33.2d). After the macroscopic
observation of the cross-section of the cup number 3 it was confirmed, that the man-
ufacturer produced a 1.5 mm thick porous layer, more than four times thicker than
the data given (Fig.33.3a). The cross-section of cup 4 did not show the division into
the surface level and the substrate. The total wall thickness of the cup was about 5
mm (Fig.33.3b). Results of potentiodynamical test showed few corrosive changes in
the form of pitting on cups made of steel. On the surfaces of porous titanium alloy
cups, in addition to numerous discolourations and residues of connective tissue, no
corrosive changes were observed (Fig.33.4).
Fig. 33.2: Results of macroscopic observations, surface of hip endoprosthesis cups:

a) 1 - residues of HAp, stereomicroscop mag. 10x, b) 2 - residues of HAp, stere-
omicroscop mag. 10x, c) 3 residuals of blood clot, stereomicroscop mag. 6x, d) 4 -
structure of bone trabeculae, stereomicroscop mag. 6x
Fig. 33.3: The cross-section of hip endoprosthesis cups: a) 3 - image of the sub-
strate with the resulting layer, stereomicroscop mag. 4.8x, b) 4 - image of cup wall,
stereomicroscop mag. 4.8x
The contents of the main chemical elements for cup 1 were: C = 0.038%, Cr
= 21.2%, Ni = 10.2%, Mo = 2.42%, Mn = 3.20%, while for cup 2: C = 0.034%,
Fig. 33.4: Sample of surface after corrosion test with example of pitting: a) visible
corrosion pits - cup 1, stereomicroscop mag. 6.3x, b) visible corrosion pits - cup 2,
stereomicroscop mag. 4x, c) surface discoloration - cup 3, stereomicroscop mag. 6x,
d) surface discoloration - cup 4, stereomicroscop mag. 1.6x
Cr = 20.9%, Ni = 9.73%, Mo = 2.39%, Mn = 3.42%. The analysis confirmed the

compliance of the chemical composition of the cups as recommended by ISO 5832 -
9. The cups 1 and 2 were made by one manufacturer but from two different batches
of material. Based on the results of the SEM (Fig.33.5), a significant proportion
of the main elements, characteristic for the examined material group was found,
that is: Cr, Ni, Mo, N. The average percentage of alloying elements for the cup
number 1 was: Crav = 11.05%, Niav = 5.33%, Moav = 2.83%, Nav = 8.50%, while
for the cup number 2 it was: Crav = 17.15%, Niav = 6.08%, Moav = 2.78%, Nav
= 1.90%. In addition, high content of iron and aluminium was distinguished (for
cup 1: Fe = 29.03%, Al = 4.53%, while for cup 2: Fe = 43.48%, Al = 6.70%).
After surface topography analysis, it was found that the pore size of austenitic steel
cups, tested for width and depth, ranged from 12 μ m to 50 μ m. On the basis of
the results of the cup 3 analysis (Fig.33.6), a clear division of the material into the
substrate, the transition layer and the outer surface was stated. In each layer, the
highest percentage of titanium was observed. The mean value was Ti = 90.57%.
Additionally, the presence of the elements such as vanadium and aluminium was
observed (V = 3.60%, Al = 7.00%), which may indicate the presence of these three
elements in the alloy composition. Moreover, the presence of the elements: Ca, K,
Si, P was found; these elements originated from the tissue surrounding the cup.
In the case of the cup 4, the main ingredient of the alloy was titanium, Ti =
89.90%. Observed was also the presence of vanadium and aluminium (V = 3.15%,
Al = 6.77%). After analysing the surface of the cup 3, both width and depth, pores
(ranging in size from 57 Âţm to 487 μ m) were obtained. After analysing of the
cross-sectional view of the cup 4, it was found that the pore size, also in width and
depth, of the analysed area was 70 μ m to 218 μ m.
The obtained values of corrosion parameter are summarised in Table 33.1.
Comparing the results, the greatest differences in the value of the repassivation
potential of Ecp were noted for steel cups and the transpassivation potential Etr for
titanium alloy cups. This may be related to different steel melt used and different
thickness of the titanium alloy surface. Approximate values of the corrosive poten-
tial for steel cups were obtained. Also, the values of the polarisation resistances were
similar. The transpassivation potential Etr , recorded for titanium alloy cups, testifies
greater corrosion resistance. A linear increase in current density of 1 mA/cm2 was
Fig. 33.5: Determine the points to analyze the chemical composition: a) cup 1, b)
cup 2
Fig. 33.6: Determine the points to analyze the chemical composition: a) cup 3, b)
cup 4
Table 33.1: Results of research pitting corrosion resistance
Number of cup Ecorr , mV Eb , mV Ecp , mV Etr , mV R p , kΩ · cm2

1 -57±12 +1177±9 +313±383 - 118±45
2 -50±28 +1118±43 +35±48 - 105±16
3 -308±45 - - +1648±684 5±4.70
4 -381±12 - - +555±100 1±0.30
observed, which was achieved at +1648 mV and +555 mV. Breakdown potential
Eb values were determined in the range of +1118÷1177 mV. Greater value of the
breakdown and the repassivation potentials was obtained for cup 1.
On the basis of the hardness tests, a significant difference in the obtained results
can be found, depending on the biomaterial used and its properties. Titanium alloy
cups are characterised by higher hardness values obtained in the range of 230 HV to
231 HV. The hardness value for the cups 2, 3 and 4 vary according to the location of
the measurement. The highest hardness was recorded for the area "II" in the middle
of the implants, e.g. cup number 4 in the area "II", hardness was 360 HV, while for
the area "I" which is characteristic for lower, side parts of the implant, hardness fell
to 323 HV. This may be due to the pressure induced during anatomical limb loading.
33.4 Conclusion
On the basis of clinical data, observations and studies, it can be concluded that
lack of full stabilisation and poor adhesion of the cups to the bones could lead to
loosening and displacements of implants and, consequently, to revision operations.
Surface topography analysis showed that the austenitic steel cup pore size was less
than recommended by the authors of the works [17, 18] for optimal bone tissue over-
growth on the implant. Confirmed was the compatibility of chemical composition of
biomaterials of the cups 1 and 2 as recommended by ISO 5832 - 9Therefore, it was
finally found that the cause of lack of bone tissue overgrowth to the substrate, in the
case of Cr-Ni-Mo steel implants, could be too small porosity of the outer surface of
the implants to prevent osteointegration. In the case of cups 3 and 4, after surface
topography analysis it was found that the pore size was normal for osteointegration,
above 30 μ m. No corrosion changes occurred on their surface. As far as the tests
are concerned, it can be concluded that the lack of integration of titanium alloy cups
should not be related to their surface but may be due to the individual characteristics
of the patients.
Acknowledgements
The work has been financed from research project no BK −201/RIB2/2017(07/020/BK_17/0022).
References
1. Mróz A., Janczak M., Sulej-Chojnacka J., Rybak T., Gierzyńska âĂŞ Dolna M.: Wpływ kata ˛
ustawienia panewki endoprotezy stawu biodrowego na obciaźenie
˛ tribologiczne pary tracej
˛
"głowa - panewka", Obróbka Plastyczna Metali, 2, 141–152 (2011)
2. Andruszkiewicz A., Wróbel B., Marzec A., Koci˛ecka A.: Strategie radzenia sobie z bólem
u pacjentów ze zmianami zwyrodnieniowymi stawu biodrowego, Problemy Piel˛egniarstwa,
(2008) 16, 237–240
3. Ridan T., Ogrodzka K., Kliś A.: Post˛epowanie Rehabilitacyjne po endoprotezoplastyce stawu
biodrowego, Praktyczna fizjoterapia rehabilitacja, 43, 6–22, (2013)
4. Semenowicz J., Mroczka A., Kajzer A., Kajzer W., Koczy B., Marciniak J.: Endoprotezoplas-
tyka stawu biodrowego z uźyciem bezcementowej panewki Avantage u chorych z czynnikami
ryzyka pooperacyjnej stabilnoÅŻci endoprotezy, Ortopedia Traumatologia & Rehabilitacja,
16, 253–263 (2014)
5. Kajzer A., Kajzer W., Semenowicz J., Mroczka A.: Corrosion resistance of hip endoprosthesis
cups in the initial state and after implantation, Solid State Phenomena, 227, 523–526, (2015)
6. Golec E., Nowak S., Golec J., Sienkiewicz D., Jurczak P., Zuzek P.: Alloplastyki rewizyjne
stawu biodrowego powikłane złamaniami kości udowej âĂŞ wyniki leczenia i rehabilitacji,
Przeglad˛ Lekarski, 65, 1, 8–12, (2008)
7. Hryniewicz T.: Wst˛ep do obróbki powierzchniowej biomateriałÃşw metalowych, Koszalin,
Politechnika Koszalińska, 23 (2007)
8. Kajzer W., Chrzanowski W., Marciniak J.: Corrosion resistance of Cr-Ni-Mo steel intended
for urological stents, International Journal of Microstructure Materials Properties, 2, 188–201,
(2007)
9. Basiaga M., Walke W., Paszenda Z., Karasiński P.: Research on electrochemical properties
SiO2 layer, intended for contact with blood, deposited by sol-gel method, European Cells and
Materials, 26, 157, (2013)
10. Kajzer W., Kajzer A., Grzeszczuk O., Semenowicz J., Mroczka A.: Badanie własności fizyko-
chemicznych bezcementowych panewek stawu biodrowego, Inźynieria Biomateriałów, 132,
9–16, (2015)
11. M. Bartmański, B. Świeczko-Źurek: Project of hip joint endoprosthesis for an individual pa-
tient with materials selection, Advances In Materials Science, 15, 1 (43), 30–36 (2015)
12. Biocybernetyka i inżynieria biomedyczna, pod red.: Nał˛ecz M., Tom 4. Biomateriały,
Warszawa, Akademicka Oficyna Wydawnicza EXIT, (2003) 1–3, 105–108, 128–134, 164–
173
13. Marciniak J.: Inźynieria Biomateriałów. Zagadnienia wybrane, Gliwice, Wydawnictwo Pra-
cowni Komputerowej Jacka Skalmierskiego, (2009) 9–20, 70, 79–83
14. Kajzer A., Kajzer W., Gołombek K., Knol M., Dzielicki J., Walke W.: Corrosion resistance,
EIS and wettability of the implants made of 316 LVM steel used in chest deformation treat-
ment, Archives of Metallurgy and Materials, 61, 767–770, (2016)
15. Kiel-Jamrozik M., Szewczenko J., Basiaga M., Nowińska K.: Technological capabilities of
surface layers formation on implant made of Ti-6Al-4V ELI alloy. Acta of Bioengineering
and Biomechanics, 15, 1, 31–37 (2015)
16. Kajzer W., Kajzer A., Grygiel-Pradelok M., Zi˛ebowicz A., Zi˛ebowicz B.: Evaluation of
physicochemical properties of TiO2 layer on AISI 316 LVM stainless steel intented for urol-
ogy. Editors: Pi˛etka, E., Badura, P., Kawa, J., Wi˛eclawek, W. (Eds.) Information Technologies
in Medicine, Advances in Intelligent Systems and Computing 471, Springer International Pub-
lishing, 2, 385–398 (2016)
17. Jaegermann Z., Ślósarczyk A.: G˛esta i porowata bioceramika korundowa w zastosowaniach
medycznych, Kraków, Uczelniane Wydawnictwo Naukowo âĂŞ Dydaktyczne AGH, 27, 78,
79, (2007) (in Polish)
18. Marciniak J.: Biomateriały, Gliwice, Wydawnictwo Politechniki Ślaskiej,
˛ (2013) 15–25, 157–
159, 180–196, 335–346, 440–441
19. http://www.bbraun.ie/documents/Plasmacup.pdf, dost˛ep: 08.05.2017r
Chapter 34
The heat treatment influence on the structure
and mechanical properties of Ti6Al4V alloy
manufactured by SLM technology
Marta Kiel-Jamrozik1 , Wojciech Jamrozik2 , and Ilona Witkowska1
Abstract. Titanium alloy can be processed via advanced powder manufacturing

routes such as selective laser melting (SLM). This field is receiving increased atten-
tion from various manufacturing sectors including medical device sectors. These ad-
vanced manufacturing techniques associated advantages include: design flexibility,
reduced production cost, reduced waste, and the opportunity to more easily man-
ufacture complex or custom-shaped implants. Unfortunately, this production tech-
nique is characterized by high temperature gradients, which results in the build-up
of thermal stresses. This is the reason why heat treatments have to be used. In this
research, the effect of several heat treatments on the microstructure and mechanical
properties of Ti6Al4V alloy processed by SLM are studied.
Keywords: biomaterials, Ti6Al4V alloy, SLM, heat treatment, structure, mechani-

cal properties
34.1 Introduction
One of biomaterials currently used for implants production is titanium, and to be

more specific titanium alloys. Implants from titanium alloys have many benefits like
high strength, good fracture toughness, high corrosion resistance, and low weight.
Mechanical properties of conventional two-phase titanium alloys (like Ti6Al4V)
strongly depend on the amount, distribution, and morphology of the phase. These
properties are the reason that titanium alloys are used in different fields of medicine.
The usefulness of metallic material for surgical implants is determined by its bio-
compatibility, which is closely related with physicochemical properties of implant
surface layer [1, 2, 3, 4].
Faculty of Biomedical Engineering, Silesian University of Technology, Roosvelta str 40, 41-800
Zabrze, Poland
e-mail: marta.kiel-jamrozik@polsl.pl · Faculty of Mechanical Engineering, Silesian
University of Technology, Konarskiego str 18a, 44-100 Gliwice, Poland

https://doi.org/10.1007/978-3-319-70063-2_34
320 Marta Kiel-Jamrozik et al.
Rapid prototyping (RP) is a group of techniques, mostly 3D printing or additive

manufacturing, used to shorten production time and decrease prototype manufactur-
ing cost and allow the production of complex shapes. Thanks to RP, it is possible
to create real 3D models in a single step manufacturing process. High quality pro-
totypes can be made directly from a virtual 3D model without the use of expensive
moulding forms, milling tools, etc. RP methods create a bridge between design and
finished products. Being able to verify a design by printing a production-ready pro-
totype before investing in expensive manufacturing equipment takes the risk out of
the prototyping process. This builds confidence before making the large investments
required at the mass production level [5]. The technology of 3D printing is the future
of innovative medicine. The introduction of new and simultaneously economically
justified solutions offers a possibility to improve the quality of rehabilitation and life
of many patients. The ability to individualize the approach to a patient can lead not
only to treatment quality enhancement, but also to human life lengthening. Collabo-
ration between physician knowledge and this engineering approach can improve the
diagnosing process as well as speed up individual treatment.
One of popular RP methods is selective laser melting (SLM). It allows the man-
ufacture products (including implants) of complex shape that have desired inner
structure and surface morphology. In the SLM device an ytterbium fibre laser is used
to melt successive layers of powder (eg. Ti6Al4V) to form a product. In SLM the
powder is heated by laser radiation, then it melts and a liquid pool is created. When
the previously melted powder cools down it solidifies and the product is formed.
The product is built layer by layer. When a cross-section of a layer is ready, plat-
form with the product is lowered by layer thickness. Afterwards a new portion of
powder is deposited to build the next product layer. This process is repeated until
the whole product is made. To prevent material curling caused by the build-up of
thermal stresses during the SLM process, the part is produced on a solid substrate.
Additionally, Ti alloys are highly reactive, thus the production process has to be
conducted under an inert argon atmosphere.
The structure of the Ti6Al4V alloy obtained in SLM process differs from the one
created in conventional fabrication processes. Microstructures of SLM Ti6Al4V is
formed during direct rapid solidification and β to α transformation at very high
cooling rates is different from the plastically deformed and the heat-treated mate-
rial. In SLM acicular/lamellar α hexagonal martensitic phase is the result of ma-
terial high cooling. Moreover, the product must undergo additional post-production
heat treatments to improve ductility to get the required mechanical properties, while
Ti6Al4V products manufactured by SLM are inherently less ductile than the con-
ventionally achieved globular microstructure [6, 7, 8]. Additionally, there is a need
to reduce thermal stresses generated during the manufacturing process.
In this paper, the influence of various heat treatment conditions on the structure
and mechanical properties of the Ti6Al4V alloy produced by the SLM method is
assessed.
34 The inuence of heat treatment... 321
A powder of Ti6Al4V alloy (chemical composition according to standard PN-

EN ISO 5832) was used as a base material for the SLM process. Samples (cubes
of volume 1 cm3 ) were produced with a 30 μ m layer thickness in an argon gas at-
mosphere. In the first stage of the research, macroscopic observations of the powder
as well as samples made of Ti6Al4V alloy were carried out. A stereoscopic micro-
scope Zeiss, SteREO Discovery v.8. (zoom 3x - 80x) was used during this period.
Next observations of the powder and samples on the micro scale were performed us-
ing a scanning electron microscope (SEM) Hitachi Tabletop Miscroscope TM3030
(zoom 40x - 3000x). The diameter of the alloy powder granules was measured.
Furthermore, the chemical composition of the micro-areas of the sample surface
were identified by energy dispersive X-ray microanalysis. In order to visualize the
structure of the analysed alloy, a digestion process was applied in a solution of the
following composition: 100ml H2 O + 2ml HF (40%) + 8 ml HNO3 (ρ = 1.4g/cm3 )
according to standard ISO 20160. Identification of structure was made using a light
microscope ZEISS Observer v.s (zoom 25x - 1000x).
Heat treatment was executed in a Nabertherm furnace. Three different annealing
parameters were applied:
• temperature T = 935◦C, t = 30min, air cooling,
• temperature T = 935◦C, t = 2h, air cooling,
• temperature T = 1000◦C, t = 30min, air cooling.
A selection of homogenizing annealing temperatures were made to take into ac-
count the temperature of the Ti6Al4V phase transitions. The temperature of the heat
treatment process influences the structure of the material as well as the volume frac-
tion of the α and β phase and mechanical and physicochemical properties.
In the next step, according to PN-EN ISO 6507-1:2007 standard, Vickers hard-
ness measurements,were performed to determine the effect of heat treatment on the
mechanical properties of Ti6Al4V titanium alloy samples. An automatic Struers
DuraScan 50 hardness meter was used for this purpose. The test was carried out at
20◦C, with 10 measurements made at a distance of successive recesses of 0.5 mm.
The Vickers hardness test was performed at a load of 49 N (HV5).
Another test was to measure the surface roughness, in accordance to PN-EN
4287:1999/A1 2010 standard, using a Taylor Hobson Ametek Surtronic s128 pro-
filometer. To assess the topography of the surface, average and standard deviation
of roughness profile was determined (Ra parameter). Measurements were carried
out at 20◦C, where 10 measurements were taken, 5 in each direction (measurements
along and across the sample). The total length of the measuring section equalling l
= 4 mm.
34.3 Results
Investigating the macroscopic images of the studied samples it was found that there
is a difference in surface state, which is connected with the direction of laser beam
heating of consecutive layers of the Ti6Al4V alloy powder. The surface of side walls
showed the highest porosity, while on the upper surface there were visible scratches
related to the laser direction during the sintering process.
SEM observations made it possible to evaluate the geometry of the Ti6Al4V alloy
powder, which was the feedstock for printing. It was observed that the diameters of
the individual granules ranged from 10μ m to 41.6μ m (Fig.34.1a). Additionally, it
was found that on the lateral surface, the material was heterogeneously melted. The
laser beam falling on a sintered area did not cover all the grains so that the granules
were not fully consolidated into a homogeneous structure (Fig. 34.1b).
(a) (b)
Fig. 34.1: SEM images of a test sample: a) Ti6Al4V powder (x1000), b) Side surface
of a test specimen with visible pores (x100)
34.3.1 Microstructure of Ti6Al4V produced by SLM
In microscopic images (Fig. 34.2), it can be observed that in the case of longitudinal
metallographic section, that there are bands that indicate the next laser-sintered lay-
ers. During the SLM process very high temperature gradients are generated by the
laser beam spot. The structure obtained after the printing process is closely corre-
lated with the solidification rate and the direction of heat dissipation. Base material
in an initial state was characterized by a martensitic two-phase structure. It is indi-
cated by the martensitic α phase and there is a presence of the β phase at the grain
boundaries. Moreover, pores were found in the analysed structure. The use of vari-
ous heat treatment parameters enabled the formation of a martensitic structure with
fewer pores. The microstructure obtained for the Ti6Al4V alloy annealed at T =
1000◦C for t = 30min was characterized by the presence of the martensitic phase α ,
phase β and the globular primary phase α . In addition, Widmanstätten α structure

has been observed with α phase visible on the β -phase grain boundaries.
34.3.2 Results of the mechanical properties of SLM Ti6Al4V
The results of Vickers hardness measurements are presented in Tab 34.1. It was
noticed that the samples made from material in the initial state had the highest hard-
ness, namely 388 HV5. The lowest hardness was obtained after the annealing pro-
cess at T = 935◦C (t = 2h) - 320 HV5. Heat treatment has reduced the hardness of the
studied material. Hardness measurements of samples annealed for 30min are simi-
lar. It was found that with the same annealing time, the heat treatment temperature
had no significant effect on hardness.
Table 34.1: Vickers hardness measurements, HV5
Initial state HT-935-30min HT-935-2h HT-1000-30min
Mean 391 354 339 347

Standard deviation 13.8 7.8 14.9 7.2
34.3.3 Results of roughness measurements of Ti6Al4V alloy

fabricated by SLM
The surface roughness results of the titanium alloy samples after 3D printing are
shown in Tab. 34.2 and Tab. 34.3. Surface topography was performed on region near
to the centre of a sample, because on edges of samples piece of material falls off. It
could be associated with the manner layers creation in the manufacturing process.
This phenomenon has resulted in high surface roughness (Ra = 4.62 - 6.07 μ m).
Table 34.2: Roughness Ti6Al4V alloy of SLM made test sample in initial state
Ra, μ m
Longitudinal section Across section
Mean 4.62 6.07

Standard deviation 0.21 0.59
(a) (b)
(c) (d)
(e) (f)
(g) (h)
Fig. 34.2: Microstructure of Ti6Al4V alloy produced by SLM, x200, a), b) initial
state transverse and longitudinal, c), d) Annealing T = 935◦C, t = 30min, transverse
and longitudinal, e), f) Annealing T = 935◦C, t = 2h, transverse and longitudinal, g),
h) Annealing T = 1000◦C, t = 30min, transverse and longitudinal
Table 34.3: Roughness Ti6Al4V alloy of SLM made test samples sfter surface treat-
ment
treatment parameters
T= 935◦C T = 935◦C T = 1000◦C

t = 30min t = 2h t = 30min
Ra, μ m 0.06 0.04 0.05
In order to obtain a roughness of Ra = 0.16 μ m (most commonly used for im-

plants), treatment of samples surface consisting of grinding and mechanical polish-
ing was required. The use of surface treatment resulted in a signicant reduction in
the surface roughness of the Ti6Al4V alloy samples. The smallest roughness (Ra =
0.04 μ m) was noticed for samples annealed at T = 935◦C in t = 30min. There were
slight differences in surface roughness irrespective of the heat treatment conditions
used.
34.4 Discussion
Due to its durability coupled with the required hardness and biocompatibility, the
Ti6Al4V alloy is one of the most widely used biomaterials in medicine. In addition
to conventional methods of producing implants from a selected alloy, like casting
or applying a plastic treatment, recent research has been conducted on the use of
spatial (3D) printing technology [5, 6, 7, 8, 9, 10, 11, 12, 13, 14].
Selective laser sintering (SLM) has so far been an unclassied and non-standardized
method of producing complex objects. The process of metal powder selective sin-
tering layer by layer under the radiation of a laser beam results in objects of various
shapes, sometimes impossible to obtain by other methods. This technology has a lot
of advantages, unfortunately there are also several drawbacks. Very high tempera-
ture gradients are produced during SLM printing by the laser beam spot. The struc-
ture obtained after this process depends mostly on the material, geometric character-
istics of the metal powder, and process parameters: laser power, speed and scanning
strategy, thickness of a single layer or the presence of protective atmosphere. The
structure obtained immediately after the SLM printing process requires additional
heat treatment to form a structure with paramagnetic properties, this determines the
set of suitable mechanical properties.
Studies have shown that the Ti6Al4V alloy produced by SLM technology was
characterized by the presence of a martensitic structure composed of an α phase
and β phase phase at grain boundaries, as conrmed by studies carried out by other
research teams [6, 7, 8, 9, 10, 12, 13, 14]. The application of an annealing pro-
cess at a temperature of T = 935◦C (regardless of time) resulted in formation of a
martensitic (α + β ) structure with a lamellar nature. In the case of annealing at T

= 1000◦C for a time of t = 2h, a martensitic structure with α phase, phase β , and
a primary globular α phase was obtained. In addition, Widmanstätten α has been
observed with an α phase visible at the β -phase grain boundary for Ti6Al4V alloy
after heat treatment. The heat treatment process had a beneficial effect on reducing
the number of pores present in the structure after the sintering process. This leads to
increased homogeneity of the structure. Reducing the internal porosity is possible
through the use of hot isostatic pressing as a type of high-pressure heat treatment.
Elimination of the pores caused by this treatment will increase the mechanical prop-
erties of the Ti6Al4V alloy [11]. The use of heat treatment resulted in lower hard-
ness of Ti6Al4V alloy. Comparable hardness values were obtained by Murr [9] and
Campanelli [12]. In addition, they investigated the effect of laser power, as well as
the size of Ti6Al4V powder grains used, on the microstructure and the mechanical
properties of the titanium alloy discussed.
Performed studies prove, that application of the correct surface treatment tech-
nique can lead to improvement of material quality produced by the SLM method.
Further studies will cover a wider range of possible treatments, to find optimal com-
binations that will be suitable for straight application in real world products such as
implants.
References
1. Kiel, M., Szewczenko, J., Marciniak, J., Nowińska, K.: Electrochemical properties of Ti-
6Al-4V ELI alloy after anodization. In: Proceedings of the Third International Conference
on Information Technologies in Biomedicine. ITIB’12, Berlin, Heidelberg, Springer-Verlag
(2012) 369–378
2. Kiel, M., Szewczenko, J., Walke, W., Maricniak, J.: Application of eis method for evaluation
of physicochemical properties of modified Ti-6Al-4V ELI alloy. Przeglad ˛ Elektrotechniczny
88 (12B) (2012) 232–235
3. Kiel-Jamrozik, M., Szewczenko, J., M.Basiaga, Nowińska, K.: Technological capabilities of
surface layers formation on implants made of Ti-6Al-4V ELI alloy. Acta of Bioengineering
and Biomechanics 17 (2015) 31–37
4. Szewczenko, J., Basiaga, M., Grygiel, M., Kiel-Jamrozik, M., Kaczmarek, M.: Corrosion
resistance of Ti6Al7Nb alloy after various surface modifications. In: Corrosion and Surface
Engineering. Volume 227 of Solid State Phenomena., Trans Tech Publications (2015) 48–486
5. Chlebus, E.: Innowacyjne technologie Rapid Prototyping - Rapid Tooling w rozwoju pro-
duktu. Oficyna Wydawnicza Politechniki Wrocławskiej, Wrocław (2003)
6. Thijs, L., Verhaeghe, F., Craeghs, T., Humbeeck, J.V., Kruth, J.P.: A study of the microstruc-
tural evolution during selective laser melting of Ti-6Al-4V. Acta Materialia 58(9) (2010)
3303–3312
7. Krakhmalev, P., Fredriksson, G., Yadroitsava, I., Kazantseva, N., du Plessis, A., Yadroitsev,
I.: Deformation behavior and microstructure of Ti6Al4V manufactured by SLM. Physics
Procedia 83 (2016) 778–788
8. Vrancken, B., Thijs, L., Kruth, J.P., Humbeeck, J.V.: Heat treatment of Ti6Al4V produced
by selective laser melting: Microstructure and mechanical properties. Journal of Alloys and
Compounds 541 (2012) 177185
9. Murr, L., Quinones, S., Gaytan, S., Lopez, M., Rodela, A., Martinez, E., Hernandez, D., Mar-
tinez, E., Medina, F., Wicker, R.: Microstructure and mechanical behavior of Ti-6Al-4V pro-
duced by rapid-layer manufacturing, for biomedical applications. Journal of the Mechanical
Behavior of Biomedical Materials 2(1) (2009) 20–32
10. Song, B., Dong, S., Zhang, B., Liao, H., Coddet, C.: Effects of processing parameters on
microstructure and mechanical property of selective laser melted Ti6Al4V. Materials and
Design 35 (2012) 120–125
11. Thone, M., Leuders, S., Riemer, A., Troster, T., Richard, H.: Influence of heat-treatment on
selective laser melting products - e.g. Ti6Al4V. Solid Freeform Fabrication Proceedings 26
(2012) 492–498 Annual international solid freeform fabrication symposium by University of
Texas.
12. Campanelli, S.L., Contuzzi, N., Ludovico, A.D., Caiazzo, F., Cardaropoli, F., Sergi, V.: Manu-
facturing and characterization of Ti6Al4V lattice components manufactured by selective laser
melting. Materials 7(6) (2014) 4803–4822
13. Sidambe, A.T.: Biocompatibility of advanced manufactured titanium implants - a review.
Materials 7(12) (2014) 8168–8188
14. Ramosoeu, M., Chikwanda, H., Bolokang, A., Booysen, G., Ngonda, T.: Additive manufactur-
ing: Characterization of Ti-6Al-4V alloy intedned for biomedical application. The Southern
African Institute of Mining and Metallurgy (2010) 337–343
Chapter 35
The Funcionalization of Grade 4 Surface Used
for Blood Contacting Implants
Marcin Basiaga1 , Magdalena Antonowicz1 , Witold Walke1 , Zbigniew Paszenda1 ,

and Bogusław Zi˛ebowicz2
Abstract. The aim of this study was to evaluate selected functional qualities of
pure titanium with modified surface which could be used for implants improvement.
Modified titanium surface with various advanced engineering methods is widely
used for example for heart valve prostheses [1], implantable rotary blood pumps [2]
or surgical tools for implantation procedures. Pure titanium Grade 4 was selected
for this research purposes and it was subjected to following surface modifications:
electrochemical polishing and application of TiO2 film by ALD method. In order to
assess the usability of proposed application technology, the examination of electro-
chemical properties has been conducted. Obtained results showed that the applied
TiO2 layer in comparison to pure titanium has the positive effect on analysed prop-
erties. It can be also stated that the layer’s thickness, thereby the number of appli-
cation cycles, had an impact on the obtained results. In addition, proper conditions’
selection of surface layers deposition had a crucial influence on the final form of the
obtained layer.
Keywords: cpTi, TiO2 , ALD, electrochemical properties
35.1 Introduction
Implantation materials are expected to be biocompatible, meaning they should not

develop ignitable reactions nor induce immunological reactions. Moreover, these
materials should not be toxic, not carcinogenic. Detailed requirements for biocom-
patibility of implants and respective examination methods are concluded in ISO
10993 standard. Modern implantation materials ought to be characterized by bioac-
tivity showing the ability to react with tissues in a specific, predetermined way.
Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze, Poland · Faculty

of Mechanical Engineering, Silesian University of Technology, Gliwice, Poland e-mail: Marcin.
Basiaga@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_35
330 Marcin Basiaga et al.
Furthermore, biomaterial should be also bio-functional, fulfilling planned in vivo

functions in assumed period of time, taking into account its’ physical. Undoubt-
edly, pure titanium is a material revealing all the above mentioned criteria. Tita-
nium shows a very good bio-tolerance in tissues’ environment and originally it was
considered as a biologically neutral element [3, 4, 5]. Such resistance is connected
with rapid formation of a thin, passive oxides layer on its surface that is strongly
bounded with basic metal, preventing the direct contact between the metal and the
environment [6, 7, 8]. Titanium is characterized by a large chemical affinity to car-
bon, nitrogen and oxide, especially when they occur in atomic structure. What is
more, nitrides, oxides, cyanides and titanium carbides are characterized by great
biocompatibility. According to [9, 10] the quick reconstruction of damaged oxide
layer, after being in contact with oxygen in the solution, causes the tissue around
the implant in contact mainly with oxide layer, not metal. Nevertheless, longer clin-
ical observations showed unfavourable interaction of certain forms of released ions,
which can cause allergy. According to Martin et. al. [11] results of performed studies
indicated that certain level of physiochemical, mechanical and titanium biocompati-
bility optimization was already reached. The above mentioned results were obtained
due to proper selection of chemical and phase composition and appropriate method
of plastic and thermal treatment. Nowadays, there are constantly performed trials of
titanium surface modifications by applying different films (mostly oxide) by chem-
ical and electrochemical methods. Modification based on Atomic Layer Deposition
(ALD) method is one of the most progressive method of covering surfaces with dif-
ferent morphologies and geometries [12, 13, 14]. ALD method consists of deposit-
ing thin films, giving proper surface roughness and an appropriate set of mechanical
properties. Already performed examination misses complex characteristics of ex-
terior layers deposited by ALD method and its reference to pure titanium’s utility
properties. It is expected that detailed analysis of formed layers along with specifica-
tion of its influence on pure titanium’s utility properties will enable its optimization
in the light of still growing requirements for biomaterials. Therefore, current study
proposes to deposition of a TiO2 film on pure titanium surface by ALD method.
Presented results are an extension of the previous study conducted by authors which
dealt with a formation of physiochemical properties of implants with miniaturized
form used in cardiovascular system [15, 16, 17].
35.2.1 Preparation of samples
Samples used in the current study were obtained from technical titanium Grade
4A with a diameter of d= 14 mm and thickness g = 2 mm. All the samples were
electrochemically polished in the bathtub before the surface modification in order
to obtain the surface roughness of Ra <0.12 μ m. This surface roughness is being
35 The Funcionalization of Grade 4 Surface. . . 331
recommended for products used in the cardiovascular system. Next, such prepared
samples were coated with TiO2 layer by ALD (Atomic Layer Deposition) method.
The deposition process of ALD method was carried out with the participation of pre-
cursors: TiCl4 and H2 O at a constant temperature 200◦C) for cycles: 500 (thickness
about 25nm), 1250 (thickness about 60 nm) and 2500 (thickness about 120 nm). At
the final stage, the samples were subjected to steam sterilization in an autoclave at a
temperature T = 134◦C), pressure p = 2.1 bar for t = 12 minutes. Samples had been
prepared this way were subjected to the following tests: potentiodynamic and po-
tentiostatic, impedance. The above mentioned examination has a crucial influence
on biocompatibility evaluation of materials used in the cardiovascular system.
35.2.2 Electrochemical properties
Potentiodynamic test
The pitting corrosion resistance test was conducted using the potentiodynamic
method, which is based on the registration of polarization curves [18, 19, 20]. The
environment used for these purposes consisted of VoltaLab PGP201 potentiostat
made by Radiometer company, reference electrode (SCE type KP-113), saturated
calomel electrode, auxiliary electrode (platinum electrode type PtP-201), anode
(tested sample) and PC equipped with VoltaMaster 4 software. The test of resis-
tance to pitting corrosion was performed in accordance with the standard PN-EN
ISO 10993-15 [21] in chemical solution that simulates artificial plasma with a fol-
lowing chemical composition and concentration: CaCl2 - 0.2 g/l, NaCl - 6.8 g/l, KCl
- 0.4g/l, MgSO4 - 0,1 g/l, NaHCO3 - 2.2 g/l, Na2 HPO4 - 0.126 g/l and NaH2 PO4
- 0.026 g/l in temperature T = 37 ± 1◦C), pH = 7.0 ± 0.2. The testing started with
opening potential EOCP determination. Subsequently, the polarization curves were
recorded starting with the initial potential value, Einit = EOCP -100 mV. The potential
changed along the anode direction at the rate of 0.16 mV/s. Based on the obtained
curves, corrosion potential Ecorr and polarization resistance R p values using Stern’s
method were determined. Research were carried out for five samples from each
group.
Potentiostatic test
Evaluation of resistance to crevice corrosion made use of the potentiostatic
method [18], recording changes in current density at + 800mV potential during
15 minutes. The measurement system was identical to the one used for potentio-
dynamic tests. Research were carried out for five samples from each group.
EIS test
In order to obtain additional information about electrochemical properties of
the surface of the analysed samples, the examination with use of electrochemical
impedance spectra (EIS) was conducted [22]. Following method enabled the analy-
sis and interpretation of processes and phenomena that are proceeding on the phases
boundary: implant - tissue environment. Measurements were carried out using Auto-
Lab PGSTAT 302N measuring system equipped with a FRA2 (Frequency Response
Analyse) - (Fig. 35.1).
Fig. 35.1: Scheme of the corrosion test
The same electrode configuration was used in the examination as it was used in
the potentiodynamic measurements. Circuit’s impedance spectra were determined
during the investigation and the obtained measurements data was adjusted to the
equivalent circuit. Based on the above mentioned investigation, the value number of
resistance R and capacity C of analysed circuits were assigned. The impedance spec-
tra of examined circuit were shown as on Nyquist’s graphs for different frequency
values and on Bode’s diagrams. The voltage amplitude of the sinusoidal activation
signal was equal to 10mV. Obtained EIS spectra were interpreted after adjusting the
smallest squares method to equivalent electrical circuit. The examination using EIS
was performed in a similar way as in the case of potentiodynamic study in artificial
plasma solution. Research were carried out for five samples from each group.
35.3 Results
35.3.1 Electrochemical Properties
Potentiodynamic test
Table 35.1 and Figure 35.2 present the results of potentiodymanic tests which
investigated the pitting corrosion resistance.
Obtained results, showed that, there is no rapid increase of current density in the
whole measurement range despite of analysed version of TiO2 layer’s formation by
ALD method. This suggests that the material has a good pitting corrosion resistance.
An evidence of the above mentioned phenomenon is lack of breakdown potential Eb
and transpassivation Etr initiation. Mean values of corrosive potential for samples in
initial state were equal to Ecorr = −389mV . Whereas, deposition of TiO2 layer by
ALD method caused a positive increase of Ecorr value - table 35.1. Additionally, the
Table 35.1: Results of potentiodynamic tests - mean values
Sample Ecorr , Rp, Icorr ,

mV kΩ cm2 μ A/cm2
cpTi - initial state −389 ± 42 117 ± 23 0.22 ± 0.02

cpTi + TiO2 500 cycles −137 ± 22 3874 ± 252 0.006 ± 0.0001
cpTi + TiO2 1250 cycles −102 ± 11 854 ± 76 0.03 ± 0.001
cpTi + TiO2 2500 cycles −53 ± 8 2220 ± 189 0.01 ± 0.0002
polarization resistance value was determined with use of Stem’s method. Moreover,
it was observed that the polarization resistance increased for samples with deposited
TiO2 film in comparison to samples in initial state. It was indicated that the material
with TiO2 film deposited in 500 cycles had the best corrosion resistance. Further-
more, it was found out that with the increase of number of TiO2 layer deposition
cycles, the pitting corrosion resistance is decreasing.
In current examination a similar TiO2 layer character was observed for base made
of steel 316LVM, for which the increase of number of cycles caused the decrease of
corrosion resistance [23].
Fig. 35.2: Typical polarization curves for cpTi + TiO2
Potentiostatic test
On the basis of performed examination, it was stated that independently on the
amount of process cycles, the analysed samples have a good resistance to this type
of corrosion - Fig. 35.3.
EIS Test
Fig. 35.3: Example of potentiostatic curve drawn for cpTi after deposition of TiO2
In the next stage of this study, in order to determine the electrochemical proper-
ties, the impendency examination (EIS) was carried out. The determined impedance
spectra do not indicate diversified corrosion processes kinetics that proceeds in
cpTi+TiO2 -artificial plasma - Fig. 39.3, 39.4, 39.5.
Fig. 35.4: Impedance spectra for sample cpTi + TiO2 (500 cycles), Nyquist’s graph
and Bode’s diagram
Impedance spectra for the analysed samples were interpreted by comparing them
to equivalent circuit, which indicates the appearance of complex oxide layer. The
oxide layer is composed of two sublayers: internal compacted and external porous,
that are mainly made of titanium oxide TiO2 - Fig. 38.3.
Symbols presented in Fig. 38.3 show respectively: Rs - artificial plasma resis-
tance, C pore - the capacity of pore layer with big surface development, R pore - pore
layer resistance, CPEdl - capacity of double layer, Rct - the charge carry resistance
on the phase boundary. This equivalent circuit (35.1) can be described by the follow-
ing dependence defining its resultant impendency .
1 1
Z = Rs + + (35.1)
1
R pore +Y ( jω )n 1
Rct +Y ( jω )n
Fig. 35.5: Impedance spectra for sample TiO2 (1250 cycles), Nyquist’s graph and
Bode’s diagram
Fig. 35.6: Impedance spectra for sample TiO2 (2500 cycles): Nyquist’s graph and-
Bode’s diagram
Based on the equivalent schemes, the characteristic magnitudes describing cor-

rosion resistance for diversified number of cycles of TiO2 layer application by ALD
method were determined - Table 35.2.
Table 35.2: Results of electrochemical impedance spectroscopy
Sample Rs , R pore , CPORE , Rct , CPEdl ,

Ω cm2 kΩ cm2 μF M Ω cm2 Y0 , Ω −1 cm−2s−n n
cpTi+ TiO2 500 cycles 77 6 8 4720 0.3981 0.92
cpTi+ TiO2 1250 cycles 76 13 11 3602 0.5965 0.87
cpTi+ TiO2 2500 cycles 75 168 13 3058 0.8169 0.81
In all Nyquist diagrams there are fragments of semicircles deformed in different

ways, located near the origin of coordinates, which then, in some cases transform
into linear dependence of imaginary impedance component (Z”) to the real one (Z’).
Fig. 35.7: Electrical model of equivalent circuit for cpTi+TiO2 layer - artificial
plasma
The results of conducted study show that, the impedance modulus is decreasing with
the frequency decrease independently of the way of pure titanium surface modifi-
cation. The layer’s thickness alteration after 500 cycles caused a bit bigger elec-
trochemical stability in comparison to layer obtained after 2500 cycles. Following
dependence can be easily noticed on Rct value, which was relatively higher after
2500 cycles. Obtained characteristics’ shapes for all samples irrespective of layer’s
thickness indicated the presence of oxide, porous layer formed as a result of artificial
plasma influence.
35.4 Conclusions
The problem stated in this study is a continuation of examination performed by au-

thors. This study deals with physicochemical of metallic biomaterials, which are
used as implants with complicated (miniaturized) geometrical features. Former au-
thors’ examination was conducted on steel 316LVM, which is used i.a. as coronary
stents, indicated that TiO2 film has a positive effect on steel’s utility properties. It
was stated that the electrochemical properties were influenced by the number of
cycles of TiO2 film deposition and by the temperature. Similar studies were also
provided for titanium Ti6Al4V and Ti6Al7Nb [24]. In this study, comparing to
316LVM steel, different results were observed. Therefore, it was proposed to con-
duct further examination on pure titanium, which is used among others as rings for
heart valves. On the basis of obtained results, the positive effect of deposited TiO2
layer on analysed utility properties in comparison to pure titanium was found. It was
also stated, that the film applied in 500 cycles process has the greatest corrosion re-
sistance, which is testified by the highest value of polarization resistance R p - table
35.1. Similar tendency was observed in the impedance studies - table 35.2. Obtained
electrochemical results showed that the layer formed this way has higher chemical
stability, thereby is the best barrier against action of artificial plasma’s environment.
References
1. Gonsior, M., Kustosz, R., Kościelniak-Ziemniak, M., Wierzchoń, T.: Biocompatible Evalua-
tion Of Biomaterials Used In The New Polish Extracorporeal Pulsatile Heart Assist Device
ReligaHeart EXT, Archives of metallurgy and materials, 60 (3), 2271 - 2278 (2015)
2. Kustosz, R., Altyntsev, I., Darłak, M., Wierzchoń, T., Tarnowski, M., Gawlikowski, M., Gon-
sior, M., Kościelniak-Ziemniak, M.:The Tin Coatings Utilisation As Blood Contact Surface
Modification In Implantable Rotary Left Ventricle Assist Device Religaheart, Archives of
Metallurgy and Materials, 60 (3), 2253-2260 (2015)
3. Roguska, A., Pisarek, M., Belcarz, A., Marcon, L., Holdynski, M., Andrzejczuk, M., Janik-
Czachor, M.: Improvement of the bio-functional properties of TiO2 nanotubes, Applied Sur-
face Science, 388 Part B, 775-785 (2016)
4. Zieliński, A., Sobieszczyk, S., Seramak, T., Serbiński, W., Świeczko-Żurek, B., Ossowska
A.: Biocompatibility and bioactivity of load-bearing metallic implants, Advances in Materials
Science 10, 4, 21-31 (2010)
5. Wang, J.-L., Liu, R.L., Majumdar, T., Mantri, S.A., Ravi, V.A., Banerjee, R., Birbilis Acta
Biomaterialia, N.: A closer look at the in vitro electrochemical characterisation of titanium
alloys for biomedical applications using in-situ methods, In Press, Corrected Proof, Available
online 16 March (2017)
6. Manjaiah, M., Laubscher, R., F.:Corrigendum to Effect of anodizing on surface integrity of
grade 4 titanium for biomedical applications, Surface and Coatings Technology, 310 (25),
263-272 (2017)
7. Sitek, R., Kaminski, J., Borysiuk, J., Matysiak, H., Kubiak, K., Kurzydlowski,
K.J.:Microstructure and properties of titanium aluminides on Ti6Al4V titanium alloy pro-
duced by chemical vapor deposition method, Intermetallics, 36, 36-44 (2013)
8. Szili, E., Kumar, S., Smart, R., Voelcker, N.: Generation of a stable surface concentration
of amino groups on silica coated onto titanium substrates by the plasma enhanced chemical
vapour deposition method, Applied Surface Science, 255, 6846-6850 (2009)
9. Li, F.,Li, J., Huang, T., Kou, H., Zhou, L.: Compression fatigue behavior and failure mecha-
nism of porous titanium for biomedical applications, Journal of the Mechanical Behavior of
Biomedical Materials, 65, 814-823 (2017)
10. Li, Y., Munir, S., Lin, J., Wen, C.: Titanium-niobium pentoxide composites for biomedical
applications, Bioactive Materials, 1 (2), 127-131 (2016)
11. Martin, E., Lanzutti, A.,Paussa, L., Guzman, L., Fedrizzi, L.: Long term performance of
atomic layer deposition coatings for corrosion protection of stainless steel, Materials and Cor-
rosion, 66 (9), 909-914 (2015)
12. Purniawan, A., French, P.J.,Pandraud, G., Sarro P.M.:TiO2 ALD nanolayer as evanescent
waveguide for biomedical sensor applications, Procedia Engineering, 5, 1131-1135 (2010)
13. Branch, B., Dubey, M., Anderson, A. S., Artyushkova, K., Baldwin, J. B., Petsev, D., Dattel-
baum A. M.: Investigating phosphonate monolayer stability on ALD oxide surfaces, Applied
Surface Science, 288 (1), 98-108 (2014)
14. Aarik, L., Arroval, T., Rammula, R., M´’n dar, H., Sammelselg, V., Aarik J.:Atomic Layer
deposition of TiO2 from TiCl4 and O3, Thin Solid Films 542, M. R. Saleem, P. Silfsten, S.
Honkanen, J. Turunen: Thin Solid Films 520, 100-107 (2013)
15. Basiaga, M., Staszuk, M., Walke, W., Opilski, Z.:Mechanical properties of ALD TiO2 layers
of stainless steel substrate, Materialwissenschaft und Werkstofftechnik, 47 (5), 512-520 (2016)
16. Basiaga, M., Jendruś, R., Walke, W.,Paszenda, Z., Kaczmarek, M ., Popczyk, M.: Influence of
Surface modification on properties of stainless steel used for implants, Archives of Metallurgy
and Materials, 60 (4), 2965-2969 (2015)
17. Karambakhsh, A., Afshar, A., Malekinejad, P.: Corrosion Resistance and Color Properties of
Anodized Ti-6Al-4V, Journal of Materials Engineering and Performance, 21, 121-127 (2012)
18. Standard: ASTM F2129 - Electrochemical Corrosion Testing of Surgical Implants (Standard
Test Method for Conducting Cyclic Potentiodynamic Polarization)
19. Marciniak, J., Szewczenko, J., Kajzer, W.: Surface Modification Of Implants For Bone
Surgery, Archives of Metallurgy and Materials, 60 (3B), 13-19 (2015)
20. Kiel, M., Szewczenko, J., Walke, W.:Application of EIS method for evaluation of physico-
chemical properties of modified Ti6Al4V ELI alloy, Przeglad ˛ Elektrotechniczny, 88 (12B),
232-235 (2012)
21. Standard: PN-EN ISO 10993-15 - Biological evaluation of medical devices

22. Krauze, A., Zi˛ebowicz, A., Marciniak J. :Corrosion resistance of intramedullary nails used
in elastic osteosynthesis of children, Journal of Materials Processing Technology, 162-163,
209-214 (2005)
23. Basiaga, M., Walke, W., Staszuk, M., Kajzer, W., Kajzer, A., Nowińska, K.:Influence of ALD
process parameters on the physical and chemical properties of the surface of vascular stents,
Archives of Civil and Mechanical Engineering, 17, 32-42 (2017)
24. Basiaga, M., Kajzer, W., Walke, W., Kajzer, A., Kaczmarek, M.: Evaluation of physicochemi-
cal properties of surface modified Ti6Al4V and Ti6Al7Nb alloys used for orthopedic implants,
Materials Science and Engineering, C 68,851-860 (2016)
Chapter 36
The influence of implantation on mechanical
degradation of the nanotubular oxide layer on
titanium screws
Katarzyna Arkusz1
Abstract. The aim of this study was to analyse the damage to the oxide layer: pas-
sive and nanotubular, covering screws made of titanium alloy Ti6Al4V ELI after
implantation. Self-cutting cervical screws were modified by anodizing in an elec-
trolyte solution of ethylene glycol with the addition of 1% DI water and 0.6% wt.
NH4 F . The microscopic and electrochemical analyses were performed by scanning
electron microscopy (SEM), electrochemical impedance spectroscopy (EIS), and
open circuit potential (OCP) measurements before and after screw implantation into
pig cervical vertebrae. Obtained results confirm the possibility of nanotubular oxide
layer (TNT) formation on Ti6Al4V ELI screws and no significant damage due to
implantation was observed for the nanotubular oxide layer.
Keywords: Titania nanotubes, Ti6Al4V ELI screw, implantation, degradation
36.1 Introduction
Titanium (Ti) and its alloys are frequently used in orthopaedic applications, due to
its good mechanical properties and high corrosion resistance. Research worldwide
has focused on surface modifications to osteointegration improvement and acceler-
ated bone remodeling on Ti surfaces [1] . Thermal spraying of coatings, nitrification,
electropolishing, chemical or physical vapour deposition, and electrochemical ox-
idation (including anodizing) are the most commonly used among the Ti surface
treatments [2].
The most widely used alloy, Ti6Al4V, reaches a yield strength almost double
that of commercially pure titanium [3]. The negative effect of Ti6Al4V implant
Biomedical Engineering Division, Department of Mechanical Engineering, University of Zielona

Gora, Zielona Gora, Poland,
e-mail: k.arkusz@ibem.uz.zgora.pl,
WWW home page: http://zib.uz.zgora.pl

https://doi.org/10.1007/978-3-319-70063-2_36
340 Katarzyna Arkusz
corrosion is the release of aluminum ions causing dysfunction in the respiratory,

bone, blood and nervous systems as well as vanadium ions causing degradation
of nerve cells and accumulation in the liver and spleen [4]. Modification of the
surface of this alloy by the formation of a nanotubular oxide layer would change
the biological properties [5, 6] and electrical surface layer by increasing electron
mobility.
Nanotubular oxide layers formed on titanium alloys are characterised by good
biocompatibility, high chemical stability, and low toxicity [7]. The thickness and
topography of the oxide layers formed on the Ti alloys depend on electrolyte com-
position, potential, and time of anodizing [8]. These factors provide the possibil-
ity of controlling the morphology of the oxide layer including biological response
by hydroxyapatite forming and osteoblast ingestion [5]. Moreover, the oxide layer
limits the focal points of corrosion releasing toxic metal into the bloodstream and
surrounding tissues.
Electrochemical formation of the nanotubular oxide layer on Ti6Al4V alloys is
problematic due to the ion adsorptions and chemical solubility of vanadium in a
solution containing chloride and fluoride ions [9]. Initially, nanotubes on Ti6Al4V
alloy were formed only in non-organic electrolytes with fluoride ion additions [10].
Obtained self-organised nanotubular oxide layers covered only the alpha-phase sur-
face regions of the Ti6Al4V alloy because of the solubility of vanadium under the
presence of fluoride. Formation of nanotubes in both phases of Ti6Al4V was suc-
cessfully carried out in an organic solution–ethylene glycol [6]. However, these nan-
otubes were formed on flat foil and no attempt was made to form the nanotubular
layer on the surface of the threaded screw.
The influence of implantation on the destruction of the nanotubular dioxide layer
was analysed only on pure titanium and no significant damage was observed [11].
This research was carried out on titanium rods whose geometry is less complicated
than the screws used in implantology. Therefore, the aim of this study was to form
the nanotubes layer on self-cutting cervical Ti6Al4V ELI screws and analyse the
electrochemical and mechanical properties before and after implantation.
36.2 Experimental details
36.2.1 Materials
Self-cutting cervical Ti6Al4V ELI screws were purchased from asCp System en-
ables post-operative dynamic adaptation (LfC, Poland). Ethylene glycol (assay
99.8%), ammonium fluoride (NH4 F), sodium chloride (NaCl) were purchased from
Sigma-Aldrich (UK) and used as supplied. All solution were prepared from high
purity reagents and distilled water.
36 K.Arkusz, TNT degradation after implantation 341
36.2.2 Fabrication of TNT on self-cutting cervical Ti6Al4V ELI

screw
Self-cutting cervical Ti6Al4V ELI screws were sonicated in acetone, ethanol and
distilled water and dried in nitrogen. The formation of nanotube oxide layers was
performed by electrochemical anodizing in ethylene glycol solution (99%) with
0.6% wt. NH4 F. The formation process consists of two stages: the first potentiody-
namic and the second potentiostatic. At the beginning the screws were polarized up
to the set potential (22 V) with scan rate 0.5 V/s and then was kept at the set potential
(22 V) for further anodizing time (20 min) in the same electrolyte. Field emission
scanning electron microscopy (FESEM, JEOL JSM-7600F) and energy-dispersive
X-ray spectroscopy (EDS, Oxford INCA) were used to investigate surface morphol-
ogy and chemical composition.
36.2.3 Electrochemical measurements
Ti6Al4V ELI screws were characterised by open circuit potential and electrochem-
ical impedance spectroscopy measurements. The above-mentioned analyses were
performed before and after screw implantation into pig cervical vertebrae according
to instructions provided by the manufacturer of the implant system. OCP measure-
ments of anodized and non-modified screws were recorded for 1200s. EIS spectra
were measured over a frequency range of 105 –0.1 Hz with an acquisition of 10
points per decade and with a signal amplitude of 10 mV. A potentiostat/galvanostat
model PGSTAT-302N from AutoLab (EcoChemie, Utrecht, The Netherlands) was
used to perform these experiments. All measurements were performed in the stan-
dard three-electrode configuration with Ti6Al4V ELI screw as the working elec-
trode, the standard silver chloride electrode (EAg/AgCl =0.222 V) as the reference
electrode and a platinum foil as the auxiliary electrode in 0.9% NaCl solution at
25 ± 2oC.
36.3.1 Surface morphology of anodized and non-modified Ti6Al4V

ELI screws
Self-cutting cervical Ti6Al4V ELI screws were anodized in nonaqueous solution

(ethylene glycol), containing fluorides due to two competitive reactions [12], ac-
cording to:
Ti + 2H2 O → TiO2 + 4H + + 4e (36.1)

TiO2 + 4H + + 6F − → TiF62− + 2H2 O (36.2)
Ethylene glycol was choosen due to the low viscous and diffusion resistance, com-
pared to other polyhydric alcohol (i.e. glycerol).
The analyzed Ti6Al4 ELI screws are hereinafter designated as: Ti6Al4V screw
for brand-new and non-modified self-cutting cervical Ti6Al4V ELI screws and
TNT/Ti6Al4V screw for anodized in ethylene glycol electrolyte self-cutting cer-
vical Ti6Al4V ELI screws.
Microscopic analysis of Ti6Al4V ELI screws (Fig. 1) were performed before and
after anodization to confirm the nanotubular layer formation on their surface. As
can be seen in Fig. 36.1C, the TNT layer of vertically aligned nanotubes with diameter
50 ± 5 nm and thickness 1000 ± 100 nm was formed on brand-passivated Ti6Al4V
ELI screws (Fig. 36.1A). SEM images show arrays of opened from the top, closed at
the bottom, and vertically oriented nanotubes. The TNT layer is not uniform and
does not cover the Ti6Al4V ELI screw completely. The Ti6Al4V ELI screw surface
consists of both the remains of passivation film and a nanotubular oxide layer formed
during the anodizing process.
SEM analysis was also performed on non-modified (Fig. 36.1B) and anodized (Fig.
36.1D) Ti6Al4V ELI screws after implantation into pig cervical vertebrae. No damage
was observed to the passive oxide layer Ti6Al4V screw after implantation (Fig.
36.1B). Smoothing of the compact oxide layer and no degradation of the TNT layer
were observed after its implantation (Fig. 36.1D), which is favourable to osteoblast
integration [2, 5]. Obtained SEM images confirm that the implantation process does
not have an influence on the homogeneity and morphology of TNT layers–height
and diameter were not changed.
36.3.2 The OCP evaluation of anodized and non-modified Ti6Al4V

ELI screws
Open circuit potential measurements were performed to indicate the thermodynamic

tendency of analysed Ti6Al4V ELI screws to electrochemical oxidation in a corro-
sive medium–0.9% NaCl. The results are shown in Fig. 36.2.
During the first period of immersion, an abrupt OCP displacement towards posi-
tive potentials was observed for non-modified Ti6Al4V ELI screws during a period
of 150s. This initial increase is related to the formation and thickening of the ox-
ide film on the screw surface, improving its corrosion protection ability [7]. How-
ever, this tendency changes, showing a decrease and reaching negative values of
OCP= −36.73 vs. Ag/AgCl mV at 1200s. This indicates the requirement of sur-
face treatment improving corrosion protection [13]. Formation of nanotubular oxide
layer on bone screw surfaces indicates ongoing increase of potential before implan-
tation, reaching a value of 4.4 mV at 1200s. It is important to point out that the
Fig. 36.1: SEM analysis of anodized and non-modified Ti6Al4V ELI screws before
and after implantation into pig vertebrae
BEFORE IMPLANTATION AFTER IMPLANTATION

Ti6Al4V screw
TNT/Ti6Al4V screw
series reaches a level from which OCP potential tends to stabilise which confirms
the protective ability of the formed oxide layer [7].
After implantation, both Ti6Al4V and TNT/Ti6Al4V screws were characterized
by highly negative open circuit potential during the first period of immersion. This
is caused by microdegradation of the oxide layer and exposing of the titanium alloy
surface. Further, significant increase of the OPC value shows the tendency towards
formation of the oxidase layer in microdamaging places and suggests ongoing oxi-
dation of the nanotubular oxide layers [6, 7, 9]. This tendency was not observed for
unmodified Ti6Al4V screw indicating stable value (-216.4 mV) during test.
Fig. 36.2: Open Circuit Potential values of anodized and non-modified Ti6Al4V ELI
screws measured in 0.9% NaCl for 1200s.
36.3.3 EIS measurements of Ti6Al4V ELI screws
The impedance spectra for anodized and non-modified Ti6Al4V screws before and
after implantation into pig cervical vertebrae are shown in Fig. 36.3 and 36.4.
The Nyquist diagram (Fig. 36.3), illustrating the impedance of the surface layer
for analyzed samples, presents large circle-shaped curves which are characteristic
of thin oxide layers [8]. The decrease of the radius of the semi-circle recorded for
TNT/Ti6Al4V screws confirms increased electrolytic resistance of the electrolyte,
which is associated with the interaction between the electrolyte and TNT layers
[6]. Implantation process decreased the semicircular diameter of capacitance loop,
which means poor corrosion resistance for Ti6Al4V screws. For the TNT/Ti6Al4V
screws before and after implantation, the real (ReZ) and imaginary impedance
(-ImZ) components are very close: ReZ = 1308Ω and −ImZ = 3273Ω for the
TNT/Ti6Al4V screw before implantation and ReZ = 1324Ω and −ImZ = 4468Ω
for the TNT/Ti6Al4V screw after implantation.
These tendencies of impedance changes are clearly shown in Bode graphs
(Fig. 36.4). The one time constant in impedance spectra seen in Fig. 36.4a indicates
that the anodized and non-modified Ti6Al4V ELI screws are built of one oxide
layer: a passivation film for the Ti6Al4V screw and a nanotubular oxide layer for
the TNT/Ti6Al4V screw [12]. The nanotubular layer formed on TNT/Ti6Al4V is
reflected by a decrease in the phase angle of impedance (−Z phase) compared to
Ti6Al4V screw. It can be seen, therefore, that implantation of TNT/Ti6Al4V screws
Fig. 36.3: Nyquist plot of anodized and non-modified Ti6Al4V ELI screws recorded
in 0.9% NaCl over the frequency range 0.1–105 Hz with amplitude 10 mV vs.
EAg/AgCl =0.222 V
causes a TNT layer homogenerity decrease by increasing the -Zphase values from
70o to 75o [14]. Decrease and shift in the phase angle of impedance after implanta-
tion confirm the degradation of oxide layer on Ti6Al4V screw.
Fig. 36.4: Bode plots of anodized Ti6Al4V ELI screws recorded in 0.9% NaCl over
the frequency range 0.1–105 Hz with amplitude 10 mV vs. EAg/AgCl =0.222 V
TNT/Ti6Al4V screw reflects a significantly lower value of impedance modu-

lus than the non-modified Ti6Al4V screw (Fig. 36.4B) indicating good resistivity
[15] and potential ability to biological substance immobilization [6]. Implanta-
tion of Ti6Al4V screw caused decrease in impedance modulus (from 57165Ω to
48148Ω ) reflecting the formation of a less protective film. After implantation of
TNT/Ti6Al4V screws, the impedance modulus value is slightly higher than before,
confirming the stability of the TNT layer and its non-degradation.
36.4 Conclusions
This work reports on the damage study of the passivation film and nanotubular ox-
ide layer formed on self-cutting cervical Ti6Al4V ELI screws during implantation
process. The formation of quite uniform and vertically oriented TiO2 nanotubes
with 50 ± 5 nm diameter and 1000 ± 100 nm length on Ti6Al4V ELI screws was
confirmed. Damage study on both cervical screw with nanotubular oxide layer and
unmodified screw was performed by implanting them into pig cervical vertebrae.
No visible changes in morphology of titania nanotubes and passivation film were
observed in microscopic analysis after implantation process, however the electro-
chemical characteristics of analysed screws indicated significant differences. The
open circuit potential (OCP) of Ti6Al4V ELI screw in the NaCl solution were
shifted in the more noble direction due to TNT layer presence protecting it against
corrosion. Although nanotubular and passive layer on the titania alloy surface have
a negative OCP value after implantation process, resulting from microdegradation
of its surface only TNT layer shows tendency to form a passive surface layer and
oxidase at damaged points/locations.
The EIS (Electrochemical Impedance Spectroscopy) tests have shown that the
nanotubular oxide layer tend to increase the corrosion resistance in NaCl media.
The results of EIS tests indicate nearly the same properties (similar modulus and
phase angle of impedance values) for ohmic resistance of the electrolyte and its
penetration through nanotube films before and after implantation. Results of the per-
formed microscopic and electrochemical analysis show that the anodized TNT layer
on Ti6Al4V ELI screw is mechanically stable for clinical application and provides
better corrosion resistance.
References
1. Kirmanidou, Y., Sidira, M., Drosou, ME., Bennani, V., Bakopoulou, A., Tsouknidas, A.,
Michailidis, N., Michalakis, K.: New Ti-alloys and surface modifications to improve the me-
chanical properties and the biological response to orthopedic and dental implants: A Review.
BioMed Res. Int. 2016, 1–21 (2016)
2. Mandracci, P., Mussano, F., Rivolo, P., Carossa, S.: Surface treatments and functional coatings
for biocompatibility improvement and bacterial adhesion reduction in dental implantology.
Coating. 6, 7, 1–22 (2016)

3. Niinomi M.: Mechanical properties of biomedical titanium alloys. Mater. Sci. Eng. 243, 231–
236 (1998)
4. Gomes, C., Moreira, L., Santos, V., Ramos, A., Lyon, J., Soares P., Santos.: Assessment of the
genetic risks of a metallic alloy used in medical implants. Genet. Mol. Biol. 34, 1, 116–121
(2011)
5. Stan, M., Memet, I., Fratila, C., Krasicka-Cydzik, E., Roman, I., Dinischiutu, A.: Effects of
titanium-based nanotube films on osteoblast behaviorin vitro. J. Biomed. Mater. Res. 103A,
48–56 (2015)
6. Kaczmarek-Pawelska, A., Krasicka-Cydzik, E.: Functionalized nanotubular oxide layer on
Ti6Al4V as a regulator and biosensor of bone tissue remodeling. ASME. 73, 2, 53–61 (2015)
7. Kiel, M., Szewczenko, J., Marciniak, J., Nowinska, K.: Electrochemical properties of Ti-6Al-
4V ELI alloy after anodization. E. PiÄŹtka and J. Kawa (Eds.): ITIB 2012, LNCS 7339,
369–378 (2012)
8. Arkusz, K., Krasicka-Cydzik, E.: Influence of thermal modification and morphology of TiO2
nanotubes on their electrochemical properties for biosensors applications. J. Nanosci. Nan-
otechnol. (to appear)
9. Mardare E., Benea L., Celis J.P.: Novel nano-TiO2 layer preparation on Ti6Al4V support alloy
and their characterization Dig. J. Nanomater. Biostruct. 3, 933–939 (2012)
10. Narayanan R., Seshadri S.K.: Phosphoric Acid anodization of Ti6Al4V- structural and corro-
sion aspects. Corros. Sci. 49, 542–558 (2007)
11. Shivaram, A., Bose, S., Bandyopadhyay, A.: Mechanical degradation of TiO2 nanotubes with
and without nanoparticulate silver coating. J. Mech. Behav. Biomed. Mater. 59, 508–18 (2016)
12. Roy, P., Berger, S., Schmuki, P.: TiO2 nanotubes: synthesis and applications. Angew. Chem.
Int. Ed. 50, 2904–2939 (2011)
13. Santana Jimenez, Y., Tejera Gil, M., Torrado Guerra, M., Baltes, L., Mirza Rosca, J.: Interpre-
tation of open circuit potential of two titanium alloys for a long time immersion in physiolog-
ical fluid. Bull. Transilv. Univ. Brasov. 2, 51 197–204 (2009)
14. Manama, N., Harunb, W., Shrib, D., Ghanib, S., Kurniawanc, T., Ismaild, M., Ibrahime, M.:
Study of corrosion in biocompatible metals for implants: A review. J. Alloy. Compd. 701,
698–715 (2017)
15. Henderson, M.: A surface perspective on self-diffusion in rutile TiO2 . Surf. Sci. 419, 2, 174–
187 (1999)
Chapter 37
The Influence of Technology on the
Physicochemical and Electrochemical Properties
of the Prosthetic Materials
Anna Zi˛ebowicz1 , Anna Woźniak1 , and Bogusław Zi˛ebowicz2
Abstract. Direct Metal Laser Sintering (DMLS) is a new technology used for get-
ting e.g. prosthetic restorations by CAD/CAM procedure. The aim of work was find-
ing an answer to question if, or the whether at all, type of manufacturing technology
affect on the physicochemical and electrochemical properties of the finished prod-
uct? The work presents the results of the studies of two Co-Cr alloys, from which the
samples were obtained in conventional casting and DMLS technology. Based on ob-
tained results it was found, that the DMLS method is a preferred one for manufactur-
ing of prosthetic devices and is an alternative way for standard casting technology.
Use of the three dimensional printing process compared to casting procedure re-
duces the time of prosthetic restorations manufacturing and improve dimensional
accuracy.
Keywords: prosthetic material, wettability, SFE, corrosion, EIS
37.1 Introduction
Currently there is a noticeable increase in patient demand for the quality of pros-
thetic restorations they use. Simultaneous technology development has resulted in
increasing use of modern CAD / CAM systems based on Rapid Prototyping (3D
printing) in dental prosthetics [1, 2, 5, 6]. In dentistry-prosthetic applications, se-
lective laser powdering processes such as Selective Laser Melting, SLS (Selective
Laser Sintering), and especially Direct Metal Laser Sintering (DSLM) are used. The
DMLS method is a variant of the SLS method, both of which have been developed
Department of Biomaterials and Medical Devices Engineering, Silesian University of Technology,

Roosevelta 40, 41-800 Zabrze Poland,
e-mail: anna.ziebowicz@polsl.pl · Department of Biomedical Materials Engineering,
Silesian University of Technology, Konarskiego 18a, 44-100 Gliwice Poland,
e-mail: boguslaw.ziebowicz@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_37
350 Anna Ziebowicz, Anna Woźniak, and Bogusáaw Ziebowicz
by EOS (Electro Optical Systems) [2, 3, 4]. Fabrication of the metal framework at
CAM stage in the DMLS procedure consists in applying to the working platform a
layer of powdered material (usually a cobalt or titanium alloy) with a blade and then
sintering it selectively with a laser beam guided in accordance with a bitmap that
is a virtual recording of the formed element. Each subsequent layer of the model is
created from the newly applied powder to the previous one after the lower working
platform has been lowered. The whole process is repeated several times, until we
have final product (Fig. 37.1). At the end of the process, unused powder surplus can
be used in the next production cycle, which reduces production costs and minimizes
waste recycling. In dental dedicated devices, an infrared laser beam is used, which
can be a CO2 laser or a modern Yb fiber laser, and the whole process takes place
in a protective gas atmosphere (the choice depends on the material being processed:
titanium alloys – in argon, cobalt alloys – in nitrogen). DMLS technology provides a
layer with a height of 0.01 to 0.08 mm and prints are characterized by high repeata-
bility and dimensional accuracy of the order of 0.02 - 0.10 mm [1, 2, 3, 4, 5, 6, 7, 8].
Fig. 37.1: DMLS prosthetic restorations (made available by the Bibus Menos)
The resulting model has a uniform construction and exhibits the mechanical
properties of the material from which it was made, without the admixtures of the
combustion products that are produced by traditional casting. Additionally use of
the three dimensional printing process compared to casting procedure reduces the
time of prosthetic restorations manufacturing. The aim of the presented paper was to
examine influence of the manufacturing technology (traditional casting and DLMS)
Co–Cr prosthetic alloy on its wettability, the surface free energy as well as on its
electrochemical behavior [1, 2, 3, 4, 5, 6, 7, 8].
37.2 Material and Method
The study material was the disks made of the Co–Cr alloy with the diameter of
14 mm. The disks were divided into two groups, depending on the manufactur-
ing method, traditional casting and laser increment sintering of powder (DMLS
method). The chemical compositions of the tested alloys are presented in Table
37.1. The samples underwent mechanical treatment, which consisted of two con-
Table 37.1: The chemical compositions of the tested CoCr alloys
Element, [%]
Material
Co Cr Mo W Si N S Fe Mn
Realloy C. 59,00 25,00 3,90 10,10 1,60 0,19 - - 0,75
EOS CoCr SP2 63,80 24,70 5,10 5,40 - - 1,00 max. 0,50 max. 1,00
secutive process: mechanical grinding with the use abrasive SiC papers of 320, 500,
800 and 1000 grit and mechanical polishing, performed with SiO2 polishing pasta.
The grinding and polishing were performed with the use of a polishing-grinding
machine by Struers.
37.2.1 Surface roughness measurements
Surface roughness measurements were performed with using contact profilometer

Surtronic 3+ by Taylor Hobson. The results of the measurements were determined
by Ra parameter. Values of average of three measurements on each sample were
given.
37.2.2 Wettability test
In order to determine the surface free energy (γs ), the wetting angle was measured.
The test stand incorporating Surftens Universal goniometer by OEG and a PC with
Surftens 4.5 software to analyze the recorded drop image. Two measure liquids were
applied: distilled water and diiodomethane. Drops of the two liquids, each 1.5μ l
volume, were placed on the surface of the samples at room temperature T = 23o C.
The values for free surface energy (SFE) and their polar and dispersion properties
for the Owens-Wendt method are given in Table 37.2.
Table 37.2: The values for SFE and their polar and apolar components for measure
liquids used in Owens – Wendt method
γL , γLd , γLP ,
Measure liquid
mJ/m2 mJ/m^{2}$ mJ/m^{2}$
Distilled water 72,8 21,8 51,0
Diiodomethane 50,8 50,8 0
37.2.3 Potentiodynamic test
In order to evaluate resistance to pitting corrosion the studies were performed with
the use of potentiodynamic method by recording of anodic polarization curves. The
test were carried out as recommended by ISO 10993 - 15 standard [11]. The test
stand comprised of the VoltaLab PGP201 potentiostat, computer with VoltaMaster
4 software and three–electrodes system, where the working electrode was repre-
sented by the test sample, the auxiliary electrode was a platinum wire (PTP–201)
and reference electrode was saturated calomel electrode (SEC KP–113). The cor-
rosion tests were started with determination of open circuit potential Eocp and then
anodic polarization curves were recorded from the starting potential Einit = Eocp –
100 mV. The potential value changed along the anodic direction at a polarization
rate of 3 mV/s. Then, when the anodic current density reached value of 1 mA/cm3 ,
polarization direction was changed. On the basis of the recorded curves character-
istic values describing to pitting corrosion were determined i.e.: corrosion potential
Ecorr (mV), potential of transpassivation Etr (mV). The value of polarization resis-
tance R p was determined with the use of Stern method and corrosion current density
was calculated from the formula icorr = 0,026/R p .
37.2.4 Microscope observation
Analysis of the surface topography after the pitting corrosion test was conducted
using the scanning electron microscope – ZEISS SUPRA 35 with type SE detector
for secondary electrons in the magnification of 20000–100000x.
37.2.5 EIS test
Electrochemical Impedance Spectroscopy EIS was conducted in order to obtain ad-

ditional information on physiochemical properties of the analyzed samples (phe-
nomena occurring at the interface: material – oral environment). Measurements
were performed with using AutoLab’s PGSTAT 302N system with a FRA2 mod-
ule(Frequency Response Analyse). The arrangement of electrodes was identical to
that used in potentiodynamic test. The research carried out in the frequency range
104 - 10−3 Hz and the amplitude of signal sinusoidal actuating tension amounted to
10 mV. During the study the impedance spectrum was determined and the obtained
data measurement were matched to equivalent electrical circuit. On such basis the
numerical values of resistance R and capacity C of the analyzed systems were de-
termined. The impedance spectra of the tested system were presented in the form
of Nyquist diagrams and as a Bode diagrams. All electrochemical tests were carried
out in artificial saliva (pH = 6,8) at the temperature 37 ± 1o C.
37.3 Results
37.3.1 Surface roughness measurements
Based on obtained results it can be concluded, that type of manufacturing technology

have significantly affected of roughness surface change, despite of the same param-
eters of surface grinding and polishing. The highest values of Ra parameter were
recorded for the cast sample (Realloy C) and mean value was 0,37 μ m, while the
lowest values for the samples made with EOS CoCr SP2 material and the obtained
mean value was 0,05 μ m.
37.3.2 Wettability test
Results of wettability and surface energy calculations are presented in Table 37.3.
The obtained results of the measurements of the water - wetting angles suggest that
together with the increase of the value of contact angle, the surface hydrophobic-
ity increases as well[8]. The contact angle average value for the cast samples was
82, 75o and pointed to the hydrophilic character of cast samples of average wetta-
bility was stated. For the EOS samples was observed hydrophobic character of the
surface, and average wetting angle was 91, 83o . The values of surface energy γ were
comparable for both groups of samples. In all the samples, one can see a large dif-
ference between the energy values of the polar and the apolar components. The high
values of the apolar components and the low values of the polar ones make it pos-
sible to conclude that these surfaces exhibit a greater affinity to the apolar groups
than to the polar ones. The obtained results of the measurements of the water - wet-
ting angles show that together with the increase of the surface roughness the surface
hydrophobicity increases as well [10] - for the cast samples with a higher surface
roughness than EOS samples, smaller contact angles were obtained. Additionally
the hydrophobic properties of surface material for the dental prosthetic restoration
are more favourable. The wettability and the surface energy are the main parameters,
that have a decisive influence on bacterial adhesion to the surface of dental materials.
The hydrophilic surface and high surface energy value favour the bacterial adhesion.
On the basis of the analysis of the surface free energy, one can state that the DMLS
samples are a better material solution for prosthetic restorations manufacturing.
Table 37.3: Results of wettability and surface energy
Contact angle, [o ] Surface energy, [mJ/m2 ]

Material
Distilled,water Diiodomethane γs γsd γsp
Realloy C. 82,75 54,25 32,82 27,83 6,59
EOS CoCr SP2 91,83 57,31 30,58 28,43 2,15
37.3.3 Potentiodynamic test
Results of potentiodynamic tests carried out to evaluate the pitting corrosion resis-
tance are presented in Fig. 37.2 While the characteristic values describing to pitting
corrosion are showed in Table 37.4. For the Realloy C the mean value of corro-
sion potential Ecorr was –327 mV. The obtained values were lower compared to
EOS CoCr SP2 material, the values of which was –234 mV. For both the examined
groups of samples, the polarization curves had a similar character. The progress
of the curves was characteristic for materials of a high corrosion resistance - no
hysteresis loops were recorded. It was determined the value of transpassivation po-
tential Etr for all the tested samples. For the Realloy the mean value was + 1011
mV and for EOS CoCr SP2 material was + 946 mV. The transpassive region for all
the samples is related to the intense oxidization of cobalt Co(III) to Co(IV) [7]. It
was stated that the value of the polarization resistance Rp for the cast samples was
equal 0,39 MΩ /cm2 and for the printing samples the mean value was 0,35 MΩ /cm2 .
Similar results of the corrosion resistance for the cast samples were obtained by the
author [9] in investigation of corrosion studies of conventional CoCrMoW alloys.
Table 37.4: Results of potentiodynamic tests
Material Ecorr , [mV] Etr , [mV] R p , [MΩ /cm2 ]

Realloy C. – 327 1011 0,39
EOS CoCr SP2 – -234 946 0,35
Fig. 37.2: Examples of the polarization curves for Realloy C and EOS CoCr SP2
materials .
37.3.4 Microscope observation
Based on observation of surface topography after the pitting corrosion test using
a scanning electron microscope (Fig. 37.3a and 3b), occurrence precipitates with
increased content of molybdenum for both materials has been observed.
Fig. 37.3: SEM image of the surface of the material after the potentiodynamic test
a) Realloy C, b) EOS CoCr SP2
Because the precipitates are areas of chemical heterogeneity of the alloy, they
can cause the corrosion process growing.
37.3.5 EIS tets
The obtained electric values determined on the basis of the recorded spectra are
shown in Table 37.5. The best fit of the modelled spectra to the impedance spectra
determined during the study in artificial saliva solution in both case is provided by
equivalent circuit composed of two parallel electric elements representing capacitive
or constant of border combined with resistance transitions and resistance at high
frequencies, which could be attributed to the resistance of the electrolyte - Fig. 37.4,
where Rs - artificial saliva solution resistance, R pore - resistance of electrolyte
in porous phase, C pore - capacity of double layer, Rct and CPEdl - electric charge
transfer resistance and capacity of oxide layer.
Table 37.5: EIS results
Eocp , Rs , R pore , CPE pore Rct , CPEdl

Material
[mV] [MΩ cm2 ] [MΩ cm2 ] Y01 [MΩ cm2 ] Y02
n n
Ω −1 cm−2 s−n Ω −1 cm−2 s−n
Realloy C. – 117 58 132 0,1336E–3 0,85 2,84 0,5951E–5 0,76
EOS CoCr SP2 – 301 68 162 0,1055E–4 0,88 2,99 0,4567E–3 0,77
Fig. 37.4: Equivalent electric cicruit for both materials
37.4 Conclusions
The proper selection of physical and chemical properties is a significant issue in the
process of adjusting the functionalities of the prosthetic devices, such as bridges,
crowns or partial dentures. It has a direct impact on its final quality and biofilm
adhesion. Continuant progress in material science, as well as growing expectations
and demands from the patients and dentists drive the efforts of manufacturing of
prosthetic materials toward improved and innovative products. On the basis of the
conducted research, it can be stated that the physicochemical properties of the pros-
thetic restorations are different depending on the type of manufacturing technology.
For the materials were stated differences in the values of roughness parameter and
contact angle. According to the obtained results, the following conclusion has been
derived, when the hydrophobic phenomena is discussed, in the context of the sur-
face of biofilm adhesion: the DMLS method is a preferred one for manufacturing of
prosthetic devices and is an alternate way for standard casting technology. The pa-
rameters of potentiodynamic tests for both materials were comparable. The results
of electrochemical impedance spectroscopy tests showed the presence of a double
layer constituting good protection of biomaterials metal against the corrosion in ar-
tificial saliva solution. Use of the three dimensional printing process compared to
casting procedure reduces the time of prosthetic restorations manufacturing and im-
prove dimensional accuracy.
References
1. Majewski S., New technology used in fabrication of fixed partial dentuers: galvanoforming,
CAD/CAM technology, titan cutting and all - ceramic systems(in polish), Protetyka Stomato-
logiczna, iss. 2, vol. 57, 2007, pp. 124 - 131
2. Vijay Venekastesh K., Widyashree Nadini V., Direct Metal Laser Sintering: A Digitised Metal
Casting Technology, The Journal of Indian Prosthodontic Society, iss. 4, vol. 13, 2013, pp. 389
- 392
3. Mangano F., Chambrone L., van Noort R., Miller C, Hatton P., Mangano C., Direct Metal Laser
Sintering Titanium Dental Implants: A Review of the Current Literature, International Jurnal
of Biomaterials, vol. 2014, 2014, pp. 1 - 11
4. Walczak M., Beer K., Surowska B., Borowicz J., The issue of using remelted CoCrMo alloys
in dental prothetics, Archives of Civil and Mechanical Engineering, iss. 2, vol. 12, 2012, pp.
171 - 177,
5. Zi˛ebowicz A., Baczkowski

˛ B.: Numerical analysis of the implant - abutment system in: Pi˛etka
E., Kawa J. (Eds.): Information Technologies in Biomedicine, LNBI 7339, pp. 341-350,
Springer-Verlag Berlin Heidelberg 2012
6. Zi˛ebowicz A., Zi˛ebowicz B., Baczkowski
˛ B.: Electrochemical behavior of materials used in
dental implantological systems. Solid State Phenomena, 2015, vol. 227, pp.447-450.
7. Bojko Ł., Ryniewicz A., Bogucki R., Pałka P., Microstructural and strength studies Co-Cr-
Mo alloy on prosthetic reconstructions in casting technology and laser sintering (in polish),
Przeglad ˛ Elektrochemiczny, iss. 5, vol. 91, 2015, pp. 29 - 32
8. Merta U., Wiśniewska G., Bacterial adhesion to dental materials - a literature review, Dental
Forum (in polish)., iss. 1, vol. 41, 2013, pp. 65 - 67
9. Augustyn - Pienia˛żek J., Łukaszczyk A., Loch B., Material and corrosion studied of conven-
tional CoCrMo alloy for frame construction casts in dental prosthetics, Engineering of Biomate-
rials, iss. 130, vol. 18, 2015, pp. 2 - 9
10. Kajzer W., Kajzer A., Grygiel - Pradelok M., Zi˛ebowicz A., Zi˛bowicz B., Evaluation of physic-
ochemical properties of TiO2Âň layer on AISI 316LVM steel intended for urology. Advances
in Intelligent System and Computing 472, Information Technologies in Medicine, Vol.2., pp.
385-398, Springer-Verlag Berlin Heidelberg 2016
11. ISO 10993-15:2000 Biological evaluation of medical devices – Part 15: Identification and
quantification of degradation products from metals and alloys
Chapter 38
Influence of Calcination Temperatures on the
Morphology and Specific Surface Area
Development of ZrO2 -CeO2 -Y2 O3 -Al2 O3
Powders Obtained via Sol-Gel Route
Damian S. Nakonieczny1 , Zbigniew K. Paszenda1 , Justyna Majewska2 , Sabina

Drewniak3 , Wojciech Bogacz4 , Magdalena Antonowicz1 , Agata Sambok1 , and
Cezary Krawczyk5
Abstract. BACKGROUND: Doped with metal oxide, zirconia is one of the most
currently used ceramics for making a wide range of all-ceramic restorations in dental
prosthetics. Its use entails very good mechanical properties and the possibility of
obtaining a desirable aesthetic effect of manufactured prostheses.
OBJECTIVE: The main purpose/aim of this study was to investigate the influ-
ence of calcination temperature on the morphology and specific surface area devel-
opment of a ZrO2 − CeO2 − Y2 O3 − Al2 O3 powder system obtained with a humid-
free sol-gel process.
METHODS: Powders were obtained by the sol-gel method in an inert gas at-
mosphere at ambient temperature using, for this purpose, zirconium n-propoxide.
Morphology was examined by scanning electron microscopy (SEM), and specific
surface area was investigated with full-adsorption isotherms in nitrogen with a
Brunauer-Emmett-Teller model (BET).
RESULTS: Depending on the calcination temperature, different morphologies
and surface areas were observed. With an increase in temperature, grain growth was
observed.
CONCLUSIONS: The influence of calcination temperature on morphology and
specific surface area of the as-obtained powders was clearly evidenced.
Keywords: zirconia, prosthetic dentistry, sol-gel, calcination, morphology, BET

ing, Silesian University of Technology, Roosevelta 40 st., Zabrze, Poland · Department of Biosen-
sors and Biomedical Signal Processing, Faculty of Biomedical Engineering, Silesian University
of Technology, Roosevelt 40 st., 41-800 Zabrze, Poland · Department of Optoelectronics, Silesian
University of Technology, B. Krzywoustego 2 st., Gliwice, Poland · Department of Chemical Engi-
neering and Process Design, Faculty of Chemistry, Silesian University of Technology, M. Strzody
7 st., Gliwice, Poland · Department of Dental Technology, Medical College of Zabrze, 3 maja 63
st., Zabrze, Poland
e-mail: Damian.Nakonieczny@polsl.pl

https://doi.org/10.1007/978-3-319-70063-2_38
360 Damian Nakonieczny et al.
38.1 Introduction
Zirconia is currently the most used ceramic for manufacturing fixed, full-ceramic
restorations in fundamentally all currently used types of prostheses, implants and
abutments [1]. The use of zirconia in dentistry and generally in biomedical engineer-
ing is controversial due to the uncontrolled phase transformation of the metastable
β , γ − ZrO2 phase to the α − ZrO2 phase, which is stable at ambient temperature.
This transformation is a diffusionless, martensitic type transformation with an in-
crease in the volume of the zirconia grains of from 4% to 6% [2, 3]. These issues
are widely discussed everywhere and to this day are the subject of numerous studies
[4, 5]. Martensitic transformation in human body fluids is associated with so-called
low temperature degradation (LTD), which is an autocatalytic process of erosion of
ceramics contained in an electrolyte and at elevated temperatures [2, 6]. Numerous
of information about the changes causing martensitic-type zirconia transformation
can be found in the work of Guo and Chevalier [8, 9, 10]. Both processes can be con-
trolled by reducing the grain size of zirconia and doping with a metal oxide stabilizer
metal, which stiffens the zirconia lattice, preventing the occurrence of uncontrolled
phase transformation [10, 11]. The critical value of grain size at which ZrO2 is not
able to undergo martensitic transformation is 0.3μ m [11]. Other factors that affect
susceptibility to LTD and the uncontrolled phase transformation are [13, 14, 15, 16]
the radius of metal ions of the added dopant, the valence of the cation, distortion
of the crystal lattice, the influence of the dopant type on the formation of liquid
solution during sintering, the value of the diffusion coefficient (DMe ) and the amor-
phousness or crystallinity of the oxide dopant. The last three variables are closely
associated with heat treatment of powders, compaction of green bodies and sinter-
ing of ceramics. Sintering is a particularly important process that affects the stabil-
ity of zirconia ceramics and resistance to degradation processes occurring in human
body fluids [2, 11, 17]. For sintering of zirconia ceramics, several methods are used,
e.g., classic pressureless sintering, microwave sintering, spark plasma sintering and
both hot and cold isostatic pressing, which allows one to obtain different results
[18, 19, 20, 21, 22]. In the present study, ZrO2 − CeO2 − Al2 O3 − Y2 O3 ceramic
powders of various chemical composition selected from previous investigations and
obtained by a one-stage sol-gel route were calcined at the chosen temperatures in the
temperature range of 55 − 850◦ C [23, 24, 25]. The effect of the calcination tempera-
ture on morphology and specific surface area (SBET ) development was investigated.
38.2.1 Material
Sample preparation was described in our previous study [25]. As a zirconia precur-
sor, we used zirconium n-propoxide (ZNP, 70% wt. in propanol, Sigma-Aldrich).
38 Inuence of Calcination Temperatures. . . 361
2-propanol was used as a solvent (PrOH, Avantor). As a ceria precursor, cerium

nitrate hexahydrate (CNH, 1 M solution in 2-propanol, Sigma-Aldrich) was used.
As a yttrium precursor, yttrium nitrate hexahydrate was used (YNH, 1 M solution
in 2-propanol, Sigma-Aldrich). As an alumina precursor, aluminium isopropoxide
(AlP, 0.5 M solution in 2-propanol, Sigma Aldrich) was used. To control the hydrol-
ysis rate, acetylacetone (AcAc, Avantor) was used as a chelating agent. Ammonia
(NH4 OH, Avantor) was used to adjust the pH. Sample characteristics and molar ra-
tios are shown in Table 38.1. All laboratory procedures and the reactant addition
sequence are presented in the diagram (Fig. 38.1).
Table 38.1: Sample characteristics and reactant molar ratios
Sample Chemical composition Calcination temperature, Molar ratios

◦C ZNP : PrOH : H2 O : AcAc
E1(1) 0.85ZrO2 0.15CeO2 550 1:2:10:0.63
E1(2) 0.85ZrO2 0.15CeO2 650 1:2:10:0.63
E1(3) 0.85ZrO2 0.15CeO2 750 1:4:10:0.63
F1(1) 0.87ZrO2 0.1CeO2 0.03Y2 O3 550 1:2:10:0.63
F1(2) 0.87ZrO2 0.1CeO2 0.03Y2 O3 650 1:2:10:0.63
F1(3) 0.87ZrO2 0.1CeO2 0.03Y2 O3 750 1:2:10:0.63
G2(1) 0.82ZrO2 0.15CeO2 0.3Al2 O3 550 1:2:10:0.63
G2(2) 0.82ZrO2 0.15CeO2 0.3Al2 O3 650 1:2:10:0.63
G2(3) 0.82ZrO2 0.15CeO2 0.3Al2 O3 750 1:2:10:0.63
38.2.2 Methods
The morphology of the dried and calcinated powders was studied by scanning elec-
tron microscopy (SEM) in an FEI INSPECT S50 microscope. SBET was investigated
with full adsorption isotherms in nitrogen with a Brunauer-Emmett-Teller model
(BET) at –196.15◦ C with an Autosorb IQ2 (Quantachrome Instruments) autosorb.
Before measurement, the sample was degassed at 105◦ C for 12 h.
38.3 Results
The most developed topography was observed for sample C2 (Fig. 38.2). The table
shows the results of measurement of grain size for all analysed powder samples
(Table 38.2). Fig. 38.2 G2(3) powder after calcination.
The results of textural research are presented in the form of plots: hysteresis ad-
sorption and desorption vary throughout the pressure range (linear isotherm accord-
ing to Langmuir) adsorption and desorption isotherms BET, BJH and de Boer. The
analysis was performed for all samples. Based on the obtained results, the following
properties were characterized:
Fig. 38.1: Laboratory procedures diagram
• type of material by IUPAC classification,

• volume and pore diameter distribution,
• estimated shape of the pores and
• SBET development.
The obtained results are presented in the table (Table 38.3).
In the figures are presented example plots of BET, BJH and de Boer graphs for
sample E1(3) (Fig. 38.3).
Analysing the data contained in the (table 38.2), the impact of temperature calci-
nation on SBET and porosity of the prepared powders can be inferred. For powders
calcined at 550◦C, the highest value of SBET was observed, falling in the range
of 1.087 − 28.744m2 g−1 . However, with the increase in calcination temperature,
the development of SBET was decreased. The lowest value of SBET of 0.928m2 g−1
Table 38.2: The measured values of aggregate and agglomerate size of the as-
calcined powders
∗ The stated value is an average size for the 4 randomly selected areas were was
selected 5 grains
∗∗ R. Caruso [24]
Sample Aggregates/Agglomerates Mean grain size∗ Classication ∗∗

μm
E1(1) aggregates/agglomerates 29.22 hard agglomerates
F1(1) aggregates/agglomerates 26.30 hard agglomerates
G2(1) agglomerates 37.74 soft agglomerates
Fig. 38.2: G2(3) powder after calcination
was noted for sample A1 (3). This trend is correlated with diameter and pore vol-
ume. Along with the increase in temperature, a reduction in pore volume was ob-
served. The lowest value of Vp (0.002cm3 g−1 ) was observed for samples F1(2), and
F1(3). Inversely proportional to the decrease in pore volume was pore diameter. The
increase in calcination temperature resulted in an increase in pore diameter. The
largest value of d p (5.644 nm) was obtained for the powder G2(3).
38.4 Discussion
All prepared powders investigated in this study were prepared by a procedure de-
scribed in previous papers [23, 24, 25]. Compared with previous powders, those
analysed in this study have a varied chemical composition and various calcination
Table 38.3: Results of textural analysis
Sample Materials type - IUPAC rec- Pores distribution Pores shape Pores diameter Pores volume SBET
ommendation
μm cm3 g− 1 m2 g− 1
E1(1) mesoporous, type IV unimodal spheric 3.066 0.010 1.087
E1(2) mesoporous, type IV unimodal spheric 3.065 0.012 1.040
E1(3) mesoporous, type IV bimodal spheric 3.381 0.006 0.928
F1(1) mesoporous, type V unimodal cubic 3.425 0.006 3.130
F1(2) mesoporous, type IV bimodal cubic 4.319 0.002 1.943
F1(3) mesoporous, type IV bimodal unregular 5.635 0.002 1.314
G2(1) mesoporous, type IV unimodal spheric 4.324 0.041 28.744
Fig. 38.3: BET plots for E1(3)
temperatures. With respect to the samples calcined at 850◦C, of all samples anal-
ysed in this study, pose F1(2) and F1(3), a reduction in the size of agglomerates and
aggregates was observed (Table 2). This phenomenon of reduction of particle size
as a function of calcination temperature could be related to diffusion grain growth
during heat treatment. Lowering the calcination temperature from 850◦C with the
same dwelling times resulted in a lowering of the rate of diffusion and limited the
grain growth. In contrast, observing changes in the size of agglomerates and ag-
gregates in the range of 550 − 750◦C for samples E1(1–3) and G2(1–3), it can be
stated that a surprising reduction of grain size occurred, and it was dependent on
increasing calcination temperature. This situation can be explained by the elimina-
tion of residual organic compounds. A tendency can also be observed for the degree
of agglomeration to increase with the increase in the mole fraction of propanol. For
powders G2, where the mole fraction of PrOH was the largest, a significant degree
of agglomeration was observed. Based on the data reported by R. Caruso et al., all
three powders from group C2 were classified as so-called soft agglomerates [24].
Fig. 38.4: BET, BJH and deBoer plots for G2(1)
Others were classified as so-called hard agglomerates. With the use of adsorption
and desorption isotherms to determine the effect of calcination temperatures on the
development of SBET , the distribution and approximate shape of the pores was inves-
tigated. SBET and total pore surface is useful information that will determine which
of the powders yield the most solid green bodies, allow perfect compacting and re-
quire lower sintering temperature to densify the structure [24]. All of the obtained
powders have been classified as mesoporous materials (pore diameter d p = 2-50 nm)
according to the convention adopted by IUPAC [26, 27]. With the increase in calci-
nation temperatures, an increase in pore diameter with a decline in total pore volume
was observed (Table 38.3). This trend was observed for all groups of powders. This
relationship also involves a reduction in SBET : with the increase in pore volume,
a reduction in SBET is observed. The largest SBET development was observed for
samples G2(1 to 3). This situation is related to the largest ratio of solvent in col-
loids from which powders have been prepared. Rapid decomposition of a thermally
stable complex of ZNP-AcAc and decomposition of propanol left behind numerous
free spaces; hence, the porosity observed in powders of this group. Textural analysis
results correlate with previous investigations about the morphology of soft and hard
agglomerates [23, 24, 25].
38.5 Conclusion
In this study, the influence of calcination temperatures on the morphology and SBET
of ZrO2 − CeO2 − Al2 O3 − Y2 O3 was investigated. In comparison with our previ-
ous studies, in which a calcination temperature of 850◦C [22, 23] was used, in this
study, a reduction of the agglomerate and aggregate grain size with decreasing cal-
cination temperature was observed, for which is responsible a reduction in the rate
of diffusion. With the increase in calcination temperature, pore diameter growth was
observed and as a result, a direct reduction in the development of SBET .
References
1. Ferrari, M., Vichi, A., Zarone, F.: Zirconia abutments and restorations: from laboratory to
clinical investigations. Dental Materials, 31,3, e73–e76 (2015)
2. Chevalier, J., Gremillard, L., Virkar, AV., Clarke, DR.: The tetragonal–monoclinic transfor-
mation in zirconia: lessons learned and future trends. Journal of American Ceramic Society,
92,9, 1901–1920 (2009)
3. Fabbri, P., Piconi, C., Burresi, E., Maganami, G., Mazzanti, F., Mingazzini, C.: Lifetime es-
timation of a zirconia–alumina composite for biomedical applications. Dental Materials, 30,
138–142 (2014)
4. Alao, A., Stoll, R., Song, X–F., Miyazaki, T., Hotta, T., Shibata, Y., Yin, L.: Surface quality of
yttrium-stabilized tetragonal zirconia polycrystal in CAD/CAM milling, sintering, polishing
and sandblasting processes. Journal of Mechanical Behavior of Biomedical Materials, 65,
102–116 (2017)
5. Tredici, IG., Sebastiani, M., Massimi, F., Bemporad, E., Resimini, A., Merlati, G., Anslemi–
Tamburni, U.: Low temperature degradation resistant nanostructured yttrium-stabilized zirco-
nia for dental applications. Ceramics International, 42,7, 8190–8197 (2016)
6. Guo, X.: Property degradation of tetragonal zirconia induced by low-temperature defect reac-
tion with water molecules. Chemistry Materials, 16, 3988–3994 (2004)
7. Guo, X., Zhang, Z.: Grain size dependent grain boundary defect structure: case of doped
zirconia. Acta Materialia, 51, 2539–2547 (2003)
8. Guo, X.: Hydrothermal degradation mechanism of tetragonal zirconia. Journal of Material
Science, 36, 3737–3744 (2001)
9. Gorban, O., Synyakina, S., Volkova, G., Gorban, S., Konstantinova, T., Lyubchik, S.: Forma-
tion of metastable tetragonal zirconia nanoparticles: Competitive influence of the dopants and
surface state. Journal of Solid State Chemistry, 232, 249–255 (2015)
10. Hallmann, L., Ulmer, P., Reusser, E., Louvel, M., Hammerle, ChHF.: Effect of dopants and
sintering temperature on microstructure and low temperature degradation of dental Y-TZP-
zirconia. Journal of The European Ceramic Society, 32, 16, 4091–4104 (2012)
11. Nakoniecznny, DS., Zi˛ebowicz, A., Paszenda, ZK., Krawczyk, C.: Trends and perspectives in
modification of zirconium oxide for a dental prosthetic applications-a review. Biocybernetics
and Biomedical Engineering, in press (2016)
12. Ugas–Carrón, R., Sittner, F., Yekehtaz, M., Flege, S., Brötz, J., Ensinger, W.: Influence of
stabilizing agents on structure and protection performance of zirconium oxide films. Surface
Coatings. Technology, 204, 2064–2067 (2010)
13. Kern, F.: Ytterbia–neodymia–costabilized TZP –Breaking the limits of strength-toughness
correlations for zirconia?. Journal of European Ceramic Society, 33, 965–973 (2013)
14. Dasari, HP., Ahn, JS., Ahn, K., Park, SY., Hong, J., Kim, H., Yoon, KJ., Son, JW., Lee, HW.,
Lee, JH.: Synthesis, sintering and conductivity behavior of ceria–doped Scandia–stabilized
zirconia. Solid State Ionics, 264, 103–109 (2014)
15. del Monte, F., Larsen, W., Mackenzie, JD.: Stabilization of tetragonal ZrO2 in ZrO2 –SiO2
binary oxides. Journal of American Ceramic Society, 83, 3, 628–634 (2000)
16. Gionea, A., Andronescu, E., Voicu, G., Bleotu, C., Surdu, V-A.: Influence of hot isostatic
pressing on ZrO2 -CeO dental ceramics properties. International Journal of Pharmaceutics,
510, 2, 439–448 (2016)
17. Vasylkiv, O., Brodianska, H., Sakka, Y., Demirsky, D.: Flash spark plasma sintering of ultra-
fine yttrium-stabilized zirconia ceramics. Scripta Materialia, 121, 32–36 (2016)
18. Surzhikov, AP., Frangulyan, TS., Ghyngazov, SA., Vasilev, IP., Chermyaskii, AV.: Ceramics
International, 42, 12, 13888–13892 (2016)
19. Ng, CK., Ramesh, S., Tan, CY., Muchta, A., Rao, Somalu, M.: Microwave sintering of ceria-
doped Scandia stabilized zirconia as an electrolyte for solid oxide fuel cell. International Jour-
nal of Hydrogen Energy, 41,32, 14184–14190 (2016)
20. Scherrer, SS., Cattani-Lorente, M., Yoon, S., Karvonen, L., Pokrant, S., Rothburst, F., Kuebler,
J.: Post-hot isostatic pressing: A Healing treatment for process related defects and laboratory
grinding damage of dental zirconia?. Dental Materials, 29, 9, e180–e190 (2013)
21. Nakonieczny, DS., Paszenda, ZK., Drewniak, S., Radko, T., Lis, M.: ZrO2 − CeO2 ceramic
powders obtained from a sol-gel process using acetylacetone as a chelating agent for potential
application in prosthetic dentistry. Acta of Bioengineering and Biomechanics, 2016, 3, 18),
DOI: 10.5277/ABB-00476-2015-3 (2016)
22. Nakonieczny, DS., Paszenda, ZK., Basiaga, M., Drewniak, S., Bogacz, W., Podwórny, J.,
Radko, T.: Phase composition and morphology characteristics of ceria-stabilized zirconia
powders obtained via sol-gel method with various pH conditions. Acta of Bioengineering and
Biomechanics, 2, 19, article in press (2017)
23. Nakonieczny, DS., Paszenda, ZK., Drewniak, S., Bogacz, W., Radko, T., Krawczyk, C.: Exper-
imental investigation of particle size distribution and morphology of alumina-yttrium-ceria-
zirconia powders obtained via sol-gel route. Journal of Materials Science, unpublished yet
material.
24. Mamana, N., Díaz-Parralejo, A., Ortiz, AL., Sanchez-Bajo, F., Caruso, R.: Ceramics Interna-
tional, 40, 6421–6426 (2014)
25. Sing, KSW.: Advances in Colloid and Interface Science. 76–77, 3–11 (1998)
26. IUPAC Recommendations. Pure and Applied Chemistry, 57, 603 (1985)
27. IUPAC Recommendations. Pure and Applied Chemistry, 66, 1739 (1994)
Chapter 39
The method of fretting wear assessment with the
application of 3D laser measuring microscope
Jarosław Sidun1 and Jan Ryszard Dabrowski

˛ 1
Abstract. Degradation processes of implant materials have a significant effect on

the reactions taking place around them. Processes related to mechanical wear, cor-
rosion and tribological wear can be distinguished here. The phenomenon of fretting
has a particularly significant impact on changes around the implant. The fretting
process has a mechanism specific to it, based on simultaneous joining and inten-
sive oxidation. This paper presents a new method for fretting wear assessment was
developed, based on the application of an Olympus Lext OLS4000 confocal laser
scanning microscope in tests. The main advantage of the described method is that
measurements can be performed in three dimensions at very high resolution. How-
ever the greatest innovation of this method of fretting wear testing is the capability
of inputting digital threshold values of a given sample’s profile. This allows for easy
differentiation between material depleted and accrued as a result of fretting.
Keywords: Fretting wear, laser scanning microscopy, wear assessment
39.1 Introduction
Ensuring the proper lifetime of implants in a tissue environment is a very important

engineering and clinical problem. The dominant processes of biomaterials’ degrada-
tion include: corrosion, mechanical damage and tribological wear, particularly un-
der so-called fretting conditions. Fretting processes, i.e. friction between two parts
nominally permanently joined due to micro-displacements, are particularly danger-
ous and the least known. These processes intensify in an aggressive biological envi-
ronment with significant involvement of so-called biofilm.
Fretting is defined as a complex process of relative, oscillatory micro-displacement
of surfaces in contact with each other, causing destruction of the near-surface layers
Bialystok University of Technology, Mechanical Faculty, Wiejska 45c, 15-351 Białystok,

e-mail: j.sidun@pb.edu.pl,j.dabrowski@pb.edu.pl

https://doi.org/10.1007/978-3-319-70063-2_39
370 Jarosáaw Sidun et al.
of the parts in contact [1]. Adhesive surface damage and the formation of fatigue
cracks cause wear particles to form, which are later oxidized and hardened. These
particles act as an abrasive, and they grow in number as they are broken up, until
the surfaces are separated by a layer of oxide particles and wear conditions stabi-
lize. The phenomenon of adhesion may also cause material to be transferred to the
opposing surface, where intensive oxidation also takes place [3, 5]. Fretting is a pro-
cess of four overlapping primary wear mechanisms: adhesive, fatigue, abrasive and
corrosion [4]. Depending on the type and direction of movement, fretting is divided
into tangential, radial, torsional and rotary fretting [2, 19, 20, 21, 22, 23].
Wear is not uniformly distributed but rather localized in randomly scattered pits
accordingly to the state of the actual contact surface, and these pits are gradually
filled with wear products. Fatigue cracks, which reduce fatigue strength very signif-
icantly, form within these pits, particularly during the initial period of the process
[7]. The wear process occurring at the contact of two parts is difficult to observe, and
often only thorough analysis following a failure makes it possible to assess whether
the root cause was fretting. Thus, knowledge of the destruction mechanism of near-
surface layers in fretting is very important for increasing the resistance of designed
joints (friction pairs) to its action via designing, technology and materials [8, 9].
The fretting process has a mechanism specific to it, based on simultaneous
joining and intensive oxidation. This oxidation occurs due to the elevated reac-
tivity of the near-surface layer as a result of deformations and contact stresses
[15, 16, 17, 18, 19].
Material wear tests can be conducted under operating or laboratory conditions.
Laboratory tests make it possible to obtain results very similar to those that would
be obtained under actual operating conditions if the proper procedures are followed
and the proper tribological systems are applied. Work in the following fields is per-
formed in laboratories: load dynamics, modeling of contact geometry, lubrication,
temperature and pressures [3, 13, 14]. There are two types of operational tests: un-
programmed (statistical) and programmed.
Unprogrammed tests are conducted on a random group of objects operated under
actual conditions. Their main drawback is high cost due to the large number of tested
objects, long experiment time and technical difficulties. However there are also ad-
vantages to such tests, particularly the usefulness of their results, since significant
corrections are generally unnecessary. Programmed tests are more labor-consuming.
Special instrumentation must be applied to the tested element, and testing itself must
proceed according to the conditions provided for in the program. These conditions
are appropriately prepared through registration of statistical values of parameters
such as: loads, temperatures, velocities and accelerations. This makes it possible to
prepare a model of operating conditions.
Testing the wear of near-surface layers of machine parts is indispensable prior
to implementation of any device, machine or part into mass production. Of course,
selection of the measuring method and method of assessing the value of wear de-
pends on several factors. These include: the goal of the test, required accuracy of
measurement and test duration. It is required for the selected method to enable wear
measurement throughout the entire range of the Lorenz curve [6, 8, 9, 11]. In prac-
39 Fretting wear assessment with the application of 3D laser measuring microscope 371
tice, various quantitative methods are applied to assessment of tribological wear.

The following methods are distinguished: gravimetric, metric, profile graphing, ar-
tificial bases, pneumatic micrometer, dynamometric, clock, induction, tensometric,
radioactive isotope [11]. Each of the aforementioned methods has its limitations
and resolution. None of them account for the fact of oxide formation and growth at
the wear site. In certain cases of volumetric measurements of tribological wear, a
measured volume greater than the nominal value is observed. This paper proposes
a new methodology for assessment of fretting wear that accounts for the depletion
and accretion of material in the friction zone [10, 12].
39.2 Research material and methodology
A unique prototype, presented in Fig. 39.1, was built to perform fretting tests. The
friction system is a pin-on-disc system. Reciprocal motion within the range of 30–
300 μ m (regulated) is performed at a frequency of 0.5 – 10 Hz, with a pressure
range of 0Âă– 30 MPa. Tests can be performed with o without lubricants.
316L implantation steel was subjected to tests. Test samples were subjected
to grinding operations in order to achieve a surface roughness on the order of
Ra=0.4μ m. To eliminate impurities, samples were rinsed in ethanol and placed in a
sonifier for 10 minutes before every test.
Fig. 39.1: View of simulator for fretting testing and friction pair (pin and disc).
An LEXT OLS 4000 automated 3D measuring confocal microscope was used to

perform measurements and verify the developed methodology.
Wear assessment was conducted according to an original method based on mea-
surements with the application of an OLS4000 3D confocal laser scanning micro-
scope from Olympus. The proposed method of testing for assessment of fretting
wear differs significantly from standard and universally applied methods. A great
advantage it provides is imaging in three dimensions, which allows for precise mea-
surement of the surface area and volume of fretting depletions âĂŞ which was
unattainable until now with the application of classical procedures. The proposed
method employs many digital tools, enabling measurements of only areas of inter-
est, the statistics of which are computed in real time.
Thanks to the utilization of dedicated lenses, a laser with 405 nm wavelength
and a double confocal optical system, the OLS4000 model ensures achievement of
maximum effectiveness, affording reliable measurements of samples with narrow
angles.
Another significant feature of this microscope is its capacity of joining 500 pho-
tographs into a whole, which facilitates the creation of high-resolution wide-angle
images. What is more, 3D measurements can be performed on the basis of these
images.
In contrast to apparatus with a hard, needle-shaped stylus, measurement takes
place without contact, which ensures accurate roughness measurement regardless
of surface texture (for both delicate and adhesive samples) while also not damaging
the sample.
Figure 39.2 presents the simplified concept of the proposed method for fretting
wear assessment. Its essence is that is makes it possible to unequivocally differenti-
ate between fretting depletion and accrued material.
Calculated models were compared with results obtained using classical Dower
approach and also with Improved EASI Coefficients described in the paper [16].
We have determined performance of all systems by calculating for each of them
correlation coefficient (Table 33.1) and root mean squared error (Table 26.2) . All
obtained results are presented in tables below. Performance of the linear model is
also shown on the plot (Fig. 39.2), where it is compared with original ECG signal
as well as with two other EASI based signals, one generated using classical Dower
method and other generated using Improved EASI Coefficients method.
Fig. 39.2: Diagram of the method for fretting wear assessment applied in tests.
Depletions cannot be distinguished from accrued material in standard methods

such as the gravimetric or metric method, which undermines such tests to a great
extent. In the case of gravimetric methods, the sample is weighed before and after
wear, and the obtained weight is considered to be the sample weight with deple-
tions, meaning that material which may have accrued on the sample’s surface is not
accounted for. Thus, the obtained wear value may be false, and in certain cases,
greater than the initial value measured before the friction test. In metric methods, a
linear dimension of the tested element is measured before wear and after a specific
time of wear. This method also does not fully account for the volume of material
that has accrued on the sample’s surface.
The proposed method is an advanced expansion of the profiles method. It em-
ploys a dense mesh of profiles. Fretting wear on the surface is determined on the
basis of determined surface profiles. Thanks to such an approach, it is possible to
differentiate between depleted material (pits below the material’s surface layer) and
accrued material (areas above the material’s surface layer). Thanks to these consid-
erations, it was possible to determine whether material was depleted or transported
over the course of the process. Proper selection of the measuring procedure allows
for performance of wear measurements in real time. The wear traces in this part
were imaged and measured (Fig. 39.3). Software makes it possible to define thresh-
old values by means of a histogram or profiles. In the case of a histogram, there are
three possibilities:
• measurement of values above the threshold;

• measurement of values below the threshold;
• measurement of values between two thresholds.
In the case of profiles, there are four options:
• measurement of values above a defined threshold;

• measurement of values below a defined threshold;
• measurement of values between two defined thresholds;
• measurement of values between two defined thresholds.
39.3 Results
Surface profiles were determined in tests and then used as the basis for determination
of surface fretting wear. Wear was determined as depleted material (pits below the
material’s surface layer) and accrued material (areas above the material’s surface
layer). Examples of data from wear tests for the cases of depletion and accretion are
presented in fig. 39.4a, b and 39.5 accordingly.
Even a cursory analysis of the data given above reveals significant growth of the
amount of wear products in the friction zone in the absence of lubricants. The great-
est material depletion is visible only in the environment of Mucin III and Xanthan
gum, at high pressures (30 MPa).
Fig. 39.3: Microscopic image of the sample with the fretting wear on the surface
along with visible coarseness profile.
Fig. 39.4: Wear of 316L steel sample accounting for: a) fretting depletion (bottom
profile), b) material accretion (top profile).
39.4 Summary and conclusions
This paper presents a new method of fretting wear assessment with the application
of 3D confocal laser scanning microscopy. Particular emphasis has been placed on
analysis of wear assessment methods. All of the methods applied until now are
accurate to a greater or lesser extent. However their greatest flaw is the lack of 3D
Fig. 39.5: Results presenting the volumetric wear of 316L steel accounting for ma-
terial depletion/accretion.
imaging capability, no measurement digitization and no unequivocal differentiation

between material depleted or accrued over the course of the wear process.
A new method for fretting wear assessment was developed, based on the appli-
cation of Olympus Lext OLS4000 confocal laser scanning microscope. The main
advantage of the described method is that measurements can be performed in three
dimensions at very high resolution. Thanks to this, we are easily able to determine
the surface area and volume of a fretting trace of interest. Available digital tools
making it possible to examine only structures of interest and measure them in real
time also facilitate wear assessment. However the greatest innovation of this method
of fretting wear testing is the capability of inputting digital threshold values of a
given sample’s profile. This allows for easy differentiation between material de-
pleted and accrued as a result of fretting. Based on obtained results of research, the
following general conclusions can be drawn:
• the proposed method is original and has never been used before;
• it is possible to clearly distinguish between depletions and accretions of material
in the friction zone;
• thanks to the microscope’s high resolution, it is possible to quantitatively de-
scribe even subtle changes in the friction zone.
Acknowledgements
This work was supported by Bialystok University of Technology under Research

Project S/WM/1/17
References
1. Cai Z., Zhu M., Shen H., Zhou Z., Jin X.: Torsional fretting wear behaviour of 7075 aluminium
alloy in various relative humidity environments, WearÂă267, 330–339 (2009).
2. Done V., Kesavan D., Krishna M.R., Chaise T., Nelias D.: Semi analytical fretting wear sim-
ulation including wear debris. Tribology International 109, 1-9 (2017).
3. Duisabeau L., Combrade P., Forest B.: Environmental effect on fretting of metallic materials
for orthopedic implants, Wear 256, 805-816 (2004).
4. Ebara R., Fujimura M.: Fretting fatigue behavior of Ti-6Al-4V alloy under plane bending
stress and contact stress, Wear 39, 1181-1186 (2006).
5. Everitt N.M., Ding J., Bandak G., Shipway P.H., Leen S.B., Williams E.J.: Characterization
of fretting-induced wear debris for Ti-6Al-4 V, Wear 267, 283–291 (2009).
6. Gao S., Cai Z., Quan X., Zhu M., Yu H.: Comparison between radial fretting and dual-motion
fretting features of cortical bone. Wear 43, 440–446 (2010).
7. Garcin S., Fouvry S., Heredia S.: A FEM fretting map modeling: Effect of surface wear on
crack nucleation. Wear 330âĂŞ331, 145-159 (2015).
8. Gebretsadik D.W., Hardell J.: Prakash B.: Friction and wear characteristics of different Pb-free
bearing materials in mixed and boundary lubrication regimes. Wear 340-341, 63–72 (2015).
9. Geringer J., Pellier J., Cleymand F., Forest B.: Atomic force microscopy investigations on pits
and debris related to fretting-corrosion between 316L SS and PMMA. Wear 292–293, 207-217
(2012).
10. Hallab N.J., Messina C., Skipor A., Jacobs J.J.: Differences in the fretting corrosion of metal-
metal and ceramic-metal modular juctios of total hip replacements, Journal of Orthopedics
Research 22, 250-259 (2004).
11. Hebda M., Wachał A. (1980), Trybologia, Wydawnictwa Naukowo-Techniczne Warszawa,
ISBN 83-204-0043-0.
12. Heinrichs J., Olsson M., Jacobson S.: New understanding of the initiation of material transfer
and transfer layer build-up in metal forming. In situ studies in the SEM. Wear 292âĂŞ293,
61-73 (2012).
13. Kulesza E., DÄĚbrowski J.R., Sidun J., Neyman A., Mizera J.: Fretting wear of materials –
methodological aspects of research. Acta Mechanica at Automatica 6 no.3, 58-61 (2012).
14. Kumar S., Narayanan T.S.N.S., Raman S.G.S., Seshadri S.K:, Evaluation of fretting corrosion
behaviour of CP-Ti orthopaedic implant applications, Tribology International 43, 1245-1252
(2017).
15. Neyman A.: Fretting w elementach maszyn, GdaÅĎsk (2003), ISBN 83-7348-048.
16. Pellier J., Geringer J., Forest B.: Fretting-corrosion between 316L SS and PMMA: influence
of ionic strength, protein and electrochemical conditions on material wear. Application to
orthopedic implants, Wear 271, 1563-1571 (2011).
17. Pereira K., Yue T., Abdel Wahab M.: Multiscale analysis of the effect of roughness on fretting
wear. Tribology International 110, 222-231 (2017).
18. Sivakumar M., Shanadurai K.S.K., Rajeswari S., Thulasiraman V.: Failures in stainless steel
orthopaedic implant devices: A survey, Journal of Materials Science Letters 14, 351-354
(1995).
19. Tritschler B.: Forest B.: Rieu J.: Fretting corrosion of materials for orthopaedic implants: a
study of a metal/polymer contact in an artificial physiological medium, Tribology International
32, 587-596 (1999).
20. Vadiraj A., Kamaraj M.: Effect of surface treatments on fretting fatigue damage of biomedical
titanium alloys, Tribology International 40, 82-88 (2007).
21. Yu H.Y., Quan H.X., Cai Z.B., Gao S.S., Zhu M.H.: Radial fretting behavior of cortical bone
against titanium, Tribology Letters 31, 69-76 (2008).
22. Zhu M.H., Yu H.Y., Zhou Z.R: Radial fretting behaviours of dental ceramics, Tribology Inter-
national 39, 1255-1261 (2006).
23. Zhu M.H., Zhou Z.R.: On the mechanisms of various fretting wear modes, Tribology Interna-
tional 44, 1378-1388 (2011).
Chapter 40
The cell niches reproducing surface structure
Roman Major1 , Klaudia Trembecka-Wojciga1 , Jürgen Markus Lackner2 , and

Bogusław Major1
Abstract. The future and the development of science is therefore seen in interdis-
ciplinary areas such as biomedical engineering. Self-assembled structures, similar
to stem cell niches would inhibit fast division process and subsequently capture the
stem cells from the blood flow. By means of surface topography and the stiffness as
well as microstructure progenitor cells should be differentiated towards the forma-
tion of endothelial cells monolayer which effectively will inhibit activation of the
coagulation cascade. The idea of the material surface development met the interest
of the clinical institutions, which support the development of science in this area and
are waiting for scientific solutions that could contribute to the development of heart
assist systems. This would improve the effciency of the treatment of patients with
myocardial failure, supported with artificial heart assist systems. Innovative materi-
als would enable the redesign, in the post project activity, construction of ventricular
heart assist.
Keywords: Bio-inspired materials, niche-like structures, thin coatings, haemocom-

patibility, electron microscopy
40.1 Introduction
Stem cells in adult organs reside in specialized niches that regulate their prolifer-
ation and differentiation. The actual word “niche” means the medium was “live”
in vivo or in a lab or in-vitro microenvironment [1]. Stem cell niches are extraor-
dinarily complex which support many aspects of stem cell identity, including mul-
Institute of Metallurgy and Materials Science; Polish Academy of Sciences, 30-059 Cracow, Rey-
monta St.25, Poland
e-mail: r.major@imim.pl ·
Joanneum Research Forschungsges mbH, Institute of Surface Technologies and Photonics, Func-
tional Surfaces, Leobner Strasse 94, A-8712 Niklasdorf, Austria

https://doi.org/10.1007/978-3-319-70063-2_40
380 Roman Major et al.
tipotency or quiescence and provide necessary signals for their regulation [2]. The
structure of the niche varies between stem cell types, its composition ranging from
a single cell or cell type to many cells of varying cell types [3].
In addition to specialized cell types, the extracellular matrix (ECM) is a crucial
component of stem cell niche. It is believed that the stem cell niche provides a com-
plex array of physical signals, including cell-cell contacts and cell-matrix adhesions,
and biochemical signals, such as growth factors, to stem cells in a temporal and
spatial manner; the integration of both local and systemic cues in the niche guides
these cells to proliferation and fate specification [4, 5]. The structural, chemical and
mechanical properties of niche can also direct cell function [6].
Cells are highly responsive to the structural properties of their surroundings such
as topographical surface features and 3D structure. Physical stimuli such as mechan-
ical stiffness and topography are known to significantly impact stem cell behaviors;
being translated through adhesions, intracellular tension and mechano-transduction,
which can alter gene expression and thus cell fate [7]. The mechano-structural en-
vironment is the architecture in two and three dimensions as well as mechanical
forces such as stress and strain, all of which act in a non-linear but fairly constant
manner. Surface modification of polymers by the use of thin layers allows to form
new features of the material while maintaining or slightly change the physical prop-
erties of the polymer [8, 9, 10]. The main attention of the surface modification of
medically used polymers is put on improving the stability of the material in the di-
rect contact with tissue. Adhesion of the layer to the substrate is one of the most
important factors which classify material for the biomedical engineering. In order
to reconstruct the niche like structures the surface wrinkling are regarded. Reduc-
ing the overall strain energy in the layers subjected to compressive stress, so-called
surface “Wrinkling” appears.
40.2 Materials and diagnostic methods
40.2.1 Hydrogel coating deposition
The hydrogel coatings were deposited by the dip-coating method. The hydrogel
coating preparation on luminal side of the tube like element made of polyvinylchlo-
ride was carried out in 2 stages using dip-coating method. The tubes were washed
in 5 % aqueous alcohol solution and then with water. They were dried to constant
weight at 40°C . In the first stage the covered tubes were immersed in a solution of
hexane, containing 5 % ethylene glycol dimethacrylate (EGDMA) and 3 Cumene
hydroperoxide (CHP) for 5 min at 25°C . In a second step the samples were placed
in an aqueous solution containing 0.1 % FeCl2 , 1% ascorbic acid AA and 5 %
polyvinylpyrrolidone (PVP) for 15 min at 25°C . PVP is high hydrophilic, bio-
compatible polymer with low cytotoxicity. Therefore this material is an appropriate
40 The cell niches reproducing surface structure 381
one for many medical applications such as artificial pancreas, wound dressing, arti-
ficial skin, and cardiovascular devices [11, 12].
The discs were immersed in an aqueous polymer solution of 0.1 % Sodium do-
decyl sulfate (SDS) and inserted on the shaker for 5 min to remove unbound PVP.
The samples were washed in water, then transferred to phosphate buffered saline
(PBS). Incubated at 37°C overnight. Films were chemically cross-linked by ethy-
lene glycol di-methacrylate (EGDMA). The main criteria while designing the blood
contacting materials are no degradation of the substrate, bio and haemocompati-
bility, high adhesion and minimal effect on changing of the mechanical properties
of the polymeric substrate. The main challenge is to design and elaborate the pro-
tective coating which would protect the polymer from biological degradation. Thin
film materials, deposited on polymer substrates should be analyzed using highly
specialized research equipment.
40.2.2 Surface of modification of the controlled residual stresses
The glow discharge technique has been applied to modify the surface. The process
is carried out at reduced pressure. The setup for this process is similar to the tubular
flat fluorescent lamp electrodes. The system was evacuated to a vacuum (10−6 Pa)
during plasma etching (activation) and the polymerization carried out using a pulsed
plasma polymerisation of a low frequency. DC glow discharge was performed at a
frequency of 535 Hz and a cycle of about 25 %.
The surface was modified by plasma induced glow discharge in a controlled
gaseous atmosphere. The gas supplied by the cathode and anode pumped. Hexamethyld-
isiloxane (HMDSO C6 H18 OSi2 ) was used as the liquid precursor and evaporated at
25°C the pressure was measured using HMDSO gas flow controller and controlled
by manual latch vacuum. Other process gases, such as argon oxygen, C2 H2 were
controlled by mass flow meters and controlled in order to achieve a stable deposi-
tion conditions.
40.2.3 In vitro analysis of blood-matreial interaction - Impact-R

test
Haemocompatibility tests evaluate the interactions of bloodcontacting medical de-

vices or materials with blood and blood components [13]. The International Or-
ganisation for Standardisation (ISO) provides several test categories for evaluating
haemocompatibility: thrombosis coagulation, platelet haematology, and immunol-
ogy. The present work evaluates blood-material interactions under dynamic condi-
tions. The ISO has developed guidelines for testing medical materials that will be
placed in contact with circulating blood, but it has not provided exact test methods
or evaluation criteria. The authors of the work cited above presented a list of various
suggestions.
Fig. 40.1: The main assumptions of the Impact-R method.
The haemostatic mechanism is designed to stop bleeding from an injured blood

vessel [14]. The interaction of materials with the blood and the activation of the
blood clotting cascade both depend on the following elements: the surface mor-
phology, platelets, and coagulation proteins. Blood platelets are critical for vascular
haemostasis, as they activate readily upon contact with the exposed components of
the vessel wall. Blood platelets are roughly twenty times less abundant than erythro-
cytes, and the diameter of a platelet is only one-fifth of that of an erythrocyte. The
primary haemostatic function of platelets could lead to thrombosis. The most impor-
tant definition for haemocompatibility is therefore as follows: “A haemocompatible
material must not adversely interact with any blood components” [15]. Haemocom-
patibility tests are designed to detect adverse interactions between an artificial sur-
face (which may activate or destroy blood components) and blood [13]. Under aor-
tic flow conditions, as a result of strong shear forces, platelets play a crucial role in
blood-material interactions and are relevant to haemocompatibility. A commercially
available blood flow simulator, the Impact-R test, is equipped with a flow chamber,
a well in which the sample of blood to be tested is introduced, and a rotor with a
normalized roughness similar to that of the active surface (Fig. 40.1).
The test was designed for use with whole blood, and this system can only be
adapted for use with flat-shaped samples. More details are presented elsewhere
[16, 17]. Materials were analysed under arterial flow conditions. The tube-like el-
ements were analysed using assumptions similar to those applied in the Impact-R
test. Testing conditions similar to those described in [18, 19, 20, 21] were main-
tained during the experiments. A new device was designed by the authors to test
the tube-shaped elements under arterial flow conditions. This arterial flow simulator
was designed and developed as a point-of-care method for real-time testingblood-
cell function.
A method was developed to analyze the modifications to the inner surfaces of the
tube-like elements. This system was designed to test whole-blood platelet adhesion
and aggregation on a thrombogenic surface under flow conditions. The design of the
arterial flow simulator is presented in Fig. 40.2.
Fig. 40.2: The design of the arterial flow simulator.
Blood-material interactions were analyzed based on fluorescent techniques using

confocal microscopy (CLSM), flow cytometry and low-vacuum scanning electron
microscopy (ESEM). After the test, blood that had not adhered to the surface was
removed and retained for further analysis, and the cells that had adhered to the
surface were analyzed using fluorescently labelled antibodies.
40.2.4 Methods of microstructure analysis
Thin foils for transmission electron microscopy (TEM) analysis are prepared on
the QUANTA 200 3D device (using a focused gallium ion beam; technique FIB-
Focused Ion Beam) equipped with a micromanipulator OmniProbe, giving the film
exactly the places of interest (defect or phase boundary). The cross section study
is performed by the transmission electron microscopy equipped with field emission
gun. The physical surface structuring is done using plasma based techniques, like
laser ablation and ion beam milling.
Top view analysis revealed wrinkles formed on the surface (Fig. 40.3). TEM analy-
sis was used to determine the detailed construction of a titanium carbonitride coating
and its interaction with the PU substrate as well as with the hydrogel final coating.
Fig. 40.3: The main assumptions of the Impact-R method.
Fig. 40.4: The design of the arterial flow simulator.

The technique called, “bright and dark field observations” were used to determine
the structure of the coating. The analysis of the phase composition was carried out
using electron diffraction techniques as well as by high resolution image analysis.
Maintaining the stability of the chemical composition of the coatings is very impor-
tant aspect. High resolution transmission electron microscope was done to observe
the structure of the coating in the atomic scale. A cross section analysis of the mi-
crostructure was carried out for the samples with the crystallized thin buffer titanium
carbo-nitride coating (Fig. 40.4).
40.3.1 Blood-material interaction
A hydrogel layer applied on the inner side of the tubular element determines a con-
stant level MNF and IPA for PAC-1 receptor (Fig. 40.5), regardless of the coat-
ing. The second receptor, which activity was analyzed using flow cytometry was
selectin-P (P-SEL). The results of the analysis of platelets collected from above the
tested surface are shown in Fig. 40.6. The result of the analysis of P-SEL shows a
similar dependence as for the analysis of PAC-1.
Fig. 40.5: Analysis of the activation of the coagulation system, made on the basis of
the active platelet IIb IIIa receptor.
Fig. 40.6: Analysis of the activation of the coagulation system, made on the basis of
platelet active receptor P-selectin.
40.3.2 Endothelialisation
The final verification of surface modification was made in direct contact with en-
dothelial cells (Fig. 40.7). The proposed coating should allow to uptake stem cells
Fig. 40.7: Cell-matierial interaction.

from whole flowing blood. Then, stem cells should be differentiated to endothe-
lial cells in order to effectively inhibit the coagulation process. At this stage of the
work, an attempt has already been trained on the cells in vitro studies to show that
the proposed coatings have been designed correctly and might be used for in vivo
analysis.
40.4 Concluding remarks
• Wrinkling is obviously the surface topography formation mechanism for bio-

inspired thin films topography design. The idea is to use advanced materials
surface to reconstruct niche like structures for the stem cell differentiation.
• Ceramic coatings showed elastic properties. This is achieved by the dispersion
of crystalline particles uniformly in the amorphous structure. The elastic prop-
erties of the ceramic coatings are dependent on the appropriate mechanism of
thin nucleation from the gas phase.
• The coating, deposited with the high acetylene flow in the reactive chamber
demonstrated high haemocompatibile properties. On the basis of the dynamic
analysis on blood, the decreased activation of the coagulation system and the
immune response compared to the other analyzed coatings was observed. These
characteristics proves the selected appropriate parameters.
• The coating, deposited in the high acetylene flow, the same which demonstrated
high haemocompatibile properties, showed the strongest influence on the for-
mation of microparticles. Despite the low activation of platelets and leukocytes,
the number of membrane fragments is the greatest. The microparticles have an
indirect effect on the coagulation process, hence it can be assumed that in the
long-term use of this material in direct contact with the blood can lead to com-
plications.
• The hydrogel coatings are highly hydrophilic. An additional layer of water is
created on the surface that protects the surface from blood aggregation pro-
cesses. The hydrogel coatings are very highly absorbing liquids, which affects
the expansion of the material. The dynamic tests of the blood revealed a strong
hemolysis.
Acknowledgements
This publication was prepared under the project “Functional carbon based coatings
on titanium substrate, modified by laser ablation designed for the integration with
cardiac tissue and ultimately inhibit the blood clotting process” of the Polish Na-
tional Centre of Science and the principal research of the Institute of Metallurgy
and Materials Science Z-2 “Functional coating materials with special properties for
circulatory system”. No. 2016/21/N/ST8/00186
References
1. Linheng, L., Xie, T.: Stem cell niche: structure and function. Annual review of cell and devel-
opmental biology. Annu Rev Cell Dev Biol. 21, 605–31 (2005)
2. Gomez-Gaviro M. V., Lovell-Badge, R., Fernandez-Aviles, F., Lara-Pezzi, E.: The Vascular
Stem Cell Niche. J. of Cardiovasc. Trans. Res.5, 618–630 (2012)
3. Chen, S., Lewallen, M., Xie, T.: Adhesion in stem cell niche: biological roles and regulation.
Development. 140(2), 255-65 (2013)
4. Fuchs, E., Tumbar, T., Guasch, G.: Socializing with the neighbors: stem cells and their niche.
Cell 116(6), 769-778 (2004)
5. Moore, K. A., Lemischka, I. R.: Stem cells and their niches. Science, 311(5769), 1880-1885
(2006)
6. Farooque, T.M., Camp, Ch. H.Jr, Tison, Ch. K., Kumar, G., Parekh, S. H., Simon, C. G. Jr.:
Measurin stem cell dimensionality in tissue scaffolds. Biomaterials 35, 2558-2567 (2014)
7. Turner, L.A., Dalby, M. J.: Nanotopography -potential relevance in the stem cell niche. Bio-
mater. Sci. 2, 1574-1594 (2014)
8. Laha, A., Bhattacharyya, S., Krupanidhi, S.B.: Impact of microstructure on dielectric prop-
erties of Pb(Mg1/3Nb2/3)O3-PbTiO3 thin films. Materials Science and Engineering B. 106,
111-119 (2004)
9. Thornton, J. A.: High Rate Thick Film Growth - Annual Reviews Ann. Res. Mat. Sci. 7, 239
(1977)
10. Lackner, J. M.: Industrially-styled room-temperature pulsed laser deposition of titanium-based
coatings. Vacuum 78, 73-82 (2005)
11. Peppas, N. A.: Hydrogels in Medicine and Pharmacy, CRC Press (1987)
12. Hickey A.S., Peppas N.A., Mesh size and diffusive characteristics of semicrystalline poly
(vinyl alcohol) membranes prepared by freezing/thawing techniques Journal of Membrane
Science 11, 229-237 (1995)
13. Seyfert, U.T., Biehl, V., Schenk, J.: In vitro hemocompatibility testing of biomaterials accord-
ing to the ISO 109934 Biomol. Eng. 19, 91-96 (2002)
14. Anderson, J. M., Schoen, J.: in In vivo assessment of tissue compatibility, ed. B. D. Ratner, A.
S. Hoffman, F. J. Schoen and J. E. Lemons, Biomaterials Science, An introduction to materials
in medicine, Elsevier Academic Press, San Diego. 360-366 (2004)
15. Sanak, M., Jakiela B., Wegrzyn, W.: Assessment of hemocompatibility of materials with arte-
rial blood ow by platelet functional tests Bulletin of the Polish Academy of Sciences: Tech-
nical Sciences, 58(2), 317-322 (2010)
16. Lackner, J.M., Wilczek, P., Sanak, M., Jakiela, B., Stolarzewicz, B.: Kowalczuk, M., Sob-
ota, M., Maksymow, K., Spisak, M., Major, B., Functional cardio-biomaterials. Advances in
Materials Science. 11(2), 5-25 (2011)
17. Major, R.: Self-assembling surfaces of blood-contacting materials. Journal of Material Science
Materials in Medicine; Springer 24, 725-733 (2013)
18. Dardik, R., Savion, N., Gal, N., Varon, D.: Flow conditions modulate homocysteine induced
changes in the expression of endothelial cell genes associated with cell-cell interaction and
cytoskeletal rearrangement. Thrombosis and Haemostasis. 88, 1047- 1053 (2002)
19. Germanier Y., Tosatti S., Broggini N., Textor M., Buser D.: Enhanced bone appo- sition around
biofunctionalized sandblasted and acidetched titanium implant surfaces. A histomorphometric
study in miniature pigs. Clin Oral Impl Res. 17, 251-257 (2006)
20. Shenkman, B., Inbal, A., Tamarin, I., Lubetsky, A., Savion, N., Varon, D.: Di- agnosis of
thrombotic thrombocytopenic purpura based on modulation by patient plasma of normal
platelet adhesion under flow condition. Br. J. Haematol.. 120, 597-604 (2003)
21. Wang, X., Dorsam, R.T., Lauver, A., Wang, H., Barbera, F.A., Gibbs, S., Varon, D., Savion,
N., Friedman, S.M., Feuerstein, G.Z.: Comparative analysis of various platelet glycoprotein
IIb/IIIa antagonists on shear-induced platelet activation and adhesion J. Pharmacol. Exp. Ther.
Chapter 41
Efect Of Sterylization And Long-Term Exposure
To Artificial Urine On Corrosion Behaviour Of
Metallic Biomaterials With
Poly(Glikolide-Co-Kaprolactone) Coatings
Wojciech Kajzer1 , Joanna Jaworska3 , Katarzyna Jelonek3 , Janusz Szewczenko1 ,

Katarzyna Nowińska2 , and Anita Kajzer1
Abstract. Novel implants based on metallic alloys (stainless steel 316 LVM,
Ti6Al4V ELI and Ti6Al7Nb alloys) prepared by self-developed method of sur-
face modification are presented in this study. Implants were coated with biodegrad-
able and bioresorbable layer of poly(glycolide-co-caprolactone). Next, half of the
perpetrated specimens was subjected to radiation sterilization. After that, samples
were divided in to two groups: I - initial state (samples with sterilized "S" and non-
sterilized polymer "NS"; II - samples with sterilized "S90" and non-sterilized poly-
mer coating "NS90" after 90 days of exposure to artificial urine. As a result of the
conducted research, it was found that after long-term exposure the polymer degra-
dation was even and inconsiderable. However, reduction of adhesion and finally,
retraction of the coatings on the titanium alloys has been observed. Moreover, the
influence of sterilization process on the kinetics of metallic ion release to corrosion
environment for all metallic biomaterials has been observed.
Keywords: artificial urine, corrosion resistance, metallic biomaterial, polymer coat-

ings, poly(glycolide-co-caprolactone)
41.1 Introduction
Stenosis of the ureter or urethra may lead to urinary retention, the consequence
of which is a direct threat of the patient health. It may cause permanent damage
to some internal organs, especially kidneys [1]. The main reasons for stenosis of
ureters and urethra are prostatic hyperplasia, cancer of an ureter and urethra, and
narrowings as a consequence of the operation on the urinary tract [1]. In order to

ing, Silesian University of Technology, ul. Roosevelta 40, 44-800 Zabrze, Poland,
e-mail: wojciech.kajzer@polsl.pl · Institute of Applied Geology, Faculty of Mining and
Geology, Silesian University of Technology, Gliwice, Poland. · Centre of Polymer and Carbon
Materials of the Polish Academy of Sciences Zabrze, Poland

https://doi.org/10.1007/978-3-319-70063-2_41
392 Wojciech Kajzer et al.
reduce the consequences of urine retention, various procedures are used to expand
and maintain urethral or ureteral patency. One of the basic method of treatment
is the use specially designed urological catheters - urological stents. Continuous
development of techniques are increasingly demanding for materials used for this
type of implants. This applies particularly to their physical and chemical properties,
which should be tailored to the urine environment, in which urological implants are
staying.
Still unsolved, a very important issue is the problem of encrusting the surface of
these implants. It involves a partial or complete overlap of the surface of the crys-
tals of calcium phosphate in the form of hydroxyapatite and magnesium ammonium
phosphate. The gradual overgrowing of implant lumen leads to obstruction and may
stop the flow of urine. This contributes the need of removing the stent in a surgical
revision procedure and provide a new one. [2, 3, 4].In many studies on the problem
of encrustation it was observed that urinary tract infection in the presence of the
implant increases the deposition process of these crystals [5]. Degradation of the
implanted biomaterial takes place in an environment of human body fluids predom-
inantly from the operational and mechanical loads and corrosive action of the envi-
ronment [7]. Corrosion resistance is dependent on the type of metallic biomaterial
and condition, morphology and the crystal structure of its surface layer. Therefore,
improvement of functional properties of metal implants used in urology, but also in
other fields of medicine, such as interventional cardiology or orthopaedics and trau-
matology, can be obtained by mechanical and electrochemical processing and mod-
ify the surface layer of biomaterials [7, 8]. Besides modification of the surface itself
of the metal biomaterial, coatings made of polymeric biomaterials or biodegradable
polymers which release substances can be used. The purpose of these layers is to
prevent the development of inflammatory conditions caused by bacteria or fungi,
reducing the risk of incrustation. Additionally, for long-term implants, biodegrad-
able polymer coating may assist in the healing process by enabling the epithelial
tissue coverage of the stent [10, 11, 12]. At the same time, the metal structure of the
implant will provide continuous treatment of the narrowed sections of the ureter or
urethra [13].
Considering the problems, involved use of metallic biomaterial in urology and
their surface modification perspectives, in this study were presented the results of
corrosion resistance of metallic biomaterials with bioresorbable polymer coatings
in the environment simulating human urinary tract. The aim of the work was to de-
terminate the influence of long-term exposure to artificial urine and radiation steril-
ization on corrosion resistance of metallic substrate and degradation process of the
polymer coating.
In the tests, samples of austenitic stainless steel 316 LVM as well as Ti6Al4V ELI
and Ti6Al67Nb titanium alloys were used. Their chemical composition, microstruc-
ture and mechanical properties met the requirements according to standards ISO
5832-1, ISO 5832-3 and ISO 5832-11 respectively. The samples were obtained from
rods of 6 mm diameter. Prior to polymer coating the surfaces were ground with grade
600 grinding paper. Titanium alloy samples were additionally subjected to anodic
oxidation in phosphoric and sulphuric acid bath at potential 97V. Next, the polymer
coating based no poly(glycolide-co-caprolactone) - G-Cap (10/90) was applied to
the metal substrate. G-Cap was synthesized in bulk at temperature equal to 120◦ C
by the ring opening polymerization of glycolide with the use of Zirconium (IV)
acetylacetonate (Zr(acac)4) as a non-toxic initiator. Specimens were coated with the
G-Cap by dip-coating method. Metal rods were immersed into polymer (10% w/w)
solution in dichloromethane for 30 seconds and dried at air for one day and then
in a vacuum dryer. The G-Cap polymer used in the research was both developed
and deposited on metallic substrate at the Centre of Polymer and Carbon Materials,
Polish Academy of Sciences in Zabrze, Poland. The samples with polymer coatings
were divided in to two groups. One of them was subjected to a radiation steriliza-
tion (sterilized sample was marked as "S", no sterilized as "NS"). A radiation dose
of 25 kGy with an energy of 10MeV was used. Next half of the specimens ("S"
and "NS") was long-term (90 days) exposed to 0.1 dm3 of artificial urine [14] with
the following composition: NaCl - 6.17 g/dm3 , KCl - 4.75 g/dm3 , NaHPO4 - 4.35
g/dm3 , Na2 SO4 - 2.41 g/dm3 , MgSO4 - 0.99 g/dm3 , C6 H5 Na3 O7 - 0.94 g/dm3 dis-
tilled water, (pH = 5.5). The samples kept in artificial urine were marked as "S90"
and "NS90" respectively for sterilized and no sterilized ones.
The morphology of the samples before and after corrosion study was analyzed
using multimode atomic force microscopy (AFM) (MultiMode, di-Veeco, USA,
CA), and stereoscopic microscope Zeiss Stereo Discovery V8 with digital camera
MC5s with magnification 4x-24x.
Pitting corrosion tests were carried out by potentiodyamic method as recom-
mended by the PN-EN ISO 10993-15 in artificial urine solutine. The test set up con-
sisted of the VoltaLab PGP201 potentiostat, saturated calomel electrode SCE as the
reference electrode, the auxiliary electrode (platinium wire), the working electrode
(test samples) and the PC with VoltaMaster 4 software. The corrosion test started
with establishing the open circuit potential EOCP at currentless condition during the
time T= 120 min. The polarization curves were recorded starting with the potential
value Estart = EOCP - 100 mV. The applied scanning rate was equal to 3mV/s. Once
the anodic current density reached the value of 1mA/cm2 , or potential E reached
the value of +4V, the polarization direction was changed. On the basis of the ob-
tained curves the breakdown potential, Eb and the repasivation potential Ecp were
determined. Moreover the value of the corrosion potential Ecorr and the polarization
resistance R p were calculated with the use of Stern method. The potentiodynamic
test of pitting corrosion resistance was carried out for all groups of samples.
Metal ion concentration in artificial urine, following 90-day exposition of sam-
ples was measured with JY 2000 spectrometer by Yobin-Yvon with the use of in-
ductively coupled plasma-atomic emission spectrometry (ICP-AES).
Degradation process of polymer coating was conducted with the use of nuclear
magnetic resonance spectroscopy - NMR spectroscopy (Bruker-Avance II Ultra-
shield Plus spectrometr, 600 MHz). Especially change of the comonomeric unit
composition during degradation were monitored. The molar mass and molar mass
distribution of the polymers were determined by gel permeation chromatography -
GPC (Physics SP 8800 chromatograph and detectors: UV, Spectra-Physicsi differ-
ential refractometer, Shodex SE 61, Viscotek RI VE3580 detector and viscometer
Viscotek 270 Dual Detector Array).
The AFM (Fig. 41.1) micrographs of the samples before degradation show that G-
Cap coating was corrugated and homogenously distributed on the metal substrate.
The surface of the implants was smooth with roughness parameter of RMS = 137.34
nm and Ra = 64.009 nm.
Fig. 41.1: Example AFM micrographs of the G-Cap coating on sterilized Ti6Al4V
alloy (initial state - before degradation)
Macroscopic observations of specimens surface after long-term (90 days) expo-

sure to artificial urine solution (Fig. 41.2b) showed a decohesion of polymer coating
on 316 LVM stainless steel compared to its initial state before exposition. Particu-
larly, the cracking of the coating and partial disintegration with the surface of the
metallic biomaterial were observed. However, for Ti alloys, a complete delamina-
tion of the polymer coating was observed. The consequence of this was the sliding
and exposing of the metal substrate - Fig. 41.2e and Fig. 41.2f. In addition, local
thickening of the coating and air bubbles on the border of the polymer-metal sub-
strate were observed - Fig. 41.2c and Fig. 41.2d.
The results of pitting corrosion test of metallic biomaterials with poly(glycolide-
co-caprolactone) - G-Cap coatings in initial state and after long-term exposure to
artificial urine are shown in table 41.1.
The polarization curves for all types of 316 LVM stainless steel samples coated
with the polymer G-Cup coating were characterized by a hysteresis loop. Break-
down Eb and repasivation Ecp potentials have been identified. Occurring of pit-
Fig. 41.2: Examples of samples surface, 316LVM: a) NS - no sterilized, b) NS90 -

after 90 day of exposure, Ti6Al4V alloy: c) NS and d) NS90, Ti6Al7Nb alloy: e)
NS and f) NS90
Table 41.1: Results of the pitting corrosion tests
NS S NS90 S90
Ecorr , mV 18.0±4.5 91.5±4.9 -31.2±2.1 -112.7±22.8
Eb , mV +1550.5±22.1 +2286.5±269.5 +1396.0±152.6 +1410.0±43.2
316LVM
Ecp , mV -10.5±4.7 -61.2±9.8 60.0±14.1 -71.0±18.7
R p ,kΩ · cm2 26.8±3.6 208.2±48.4 0.4±0.02 0.5±0.2
Ecorr , mV 55.2±4.8 -162.9±31.6 -75.0±11.5 -358.9±30
Ti6Al4V Eb , mV - - - -
ELI Ecp , mV - - - -
R p , kΩ · cm2 55.1±9.6 60.9±5.3 89.9±24.7 2.2±0.4
Ecorr , mV -15.5±1.2 -250±49.0 -167.5±18.4 -385.0±67.5
Eb , mV - - - -
Ti6Al7Nb
Ecp , mV - - - -
R p , kΩ · cm2 191.1±3.4 14.0±3.3 47.3±17.6 12.0±1.4
ting corrosion for these samples were confirmed by macroscopic observations. Figs
41.2a and 41.2b show characteristic pitting corrosion pits.
It should be emphasized that for the samples in the initial state (not exposure to
artificial urine) the corrosion products were stopped under polymer surface - Fig.
41.2a. After long-term exposure, the copolymer coating was cracked. As a result,
partial decohesion of the coating and release of corrosion products into the elec-
trolyte occurred.
The values of repasivation potential Ecp determined on the basis of the polariza-
tion curves were similar to the corrosion potentials Ecorr for these samples. As a
result of sterilization process it was observed increase of corrosion Ecorr and break-
down Eb potentials in comparison to non-sterilized "NS" samples. Moreover, the
value of polarization resistance R p of the sterilized samples "S" were higher than
non-sterilized ones "NS" - table 41.1. Long-term exposure of "NS90" samples in ar-
tificial urine caused reduction of all parameters describing the corrosion resistance
in comparison to non-sterilized "NS" samples in the initial state. A similar trend was
observed for sterilized in initial state "S" and after long term exposure "S90" 316
LVM stainless steel samples - table 41.1.
The results of corrosion tests of titanium alloys with polymer coatings shows
resistance to pitting corrosion in whole measuring range (+4V). Neither the break-
down Eb nor transpassivation Etr potentials were observed It needs to be highlighted
that, for all titanium samples "S90" and "NS90" exposed 90 days in artificial urine,
decohesion of the polymer coating with metal substrate was observed. The process
of separation of polymer coating from metal substrate was initiated around 75 days
of exposure Figs. 41.2d and 41.2f.
Radiation sterilization has reduced the corrosion potential both for the Ti6Al4V
ELI and Ti6Al7Nb alloys substrates. In case of the sterilized "S" and no sterilized
"NS" Ti6Al4V ELI samples, the same values of polarization resistance R p were
observed - table 41.1. For the Ti6Al7Nb, it was found that the sterilization reduced
the value of polarization resistance R p in comparison to the non-sterilized samples.
A similar relationship was observed for Ti6Al4V ELI and Ti6Al7Nb alloys with
PLGA coating after long-time exposure to Ringer’s solution [15].
Results of release of metal ions to artificial urine after long-time (90 days) ex-
posure were converted into surface mass density and presented in table 41.2. The
results, for the sterilized and non-sterilized samples after exposure, showed that the
sterilization increased corrosion resistance of the metallic substrate. In particular,
a decrease in mass density of Fe, Cr, Ni and Mo ions release was observed for
316 LVM stainless steel - table 41.2. The same was observe for the titanium al-
loys (Ti6Al4V ELI and Ti6Al7Nb) for Ti, Al, V and Nb ions - table 41.2. Similar
effect of the radiation sterilization on metal ion release from metallic biomaterials
coated with the PLGA coatings after long-term exposure to Ringer’s solution were
observed by authors [15].
Table 41.2: Results of surface mass density for metal ions
316LVM Ti6Al4V Ti6Al7Nb

NS90 S90 NS90 S90 NS90 S90
Surface mass density, μ g/cm2
Fe 0.314±0.019 0.257±0.019 Ti 3.200± 0,038 2.467±0,024 Ti 2.905±0.019 2.062±0.019
Cr 3.933±0.099 3.252±0.038 Al 1.971± 0.038 1.762± 0.024 Al 1.714±0.019 1.157±0.048
Ni 1.610±0.038 1.219±0.071 V 1.124± 0.019 0.786± 0.043 V - -
Mo 0.429±0.038 0.210±0.019 Nb - - Nb 1.200±0.038 1.019±0.038
NMR analysis revealed even way of degradation of the poly(glycolide-co-caprolactone)

G-Cap coatings - the content of caproyl units in the polymeric chain remained
on the same level during 90 days. There was significant decrease of molecular
weight of PGCap (from 65 kDa to 15 kDa after 90 days), However, the disper-
sity (Mw/Mn) of the samples did not change - it was equal to 2. In the NMR spectra
of poly(glycolide-co-caprolactone), there was no signals corresponding to the short
chains of oligomers.The degradation process of the poly(glycolide-caprolactone)
coatings proceeds quite slowly during 90 days. This behaviour is connected with
the different microstructure of the chain, comparing to poly(lactide-co-glycolide):
caproyl units are characterized by greater hydrophobicity and therefore have a
slower hydrolysis rate.
41.4 Conclusion
On the basis of obtained results, it was shown that radiation sterilized poly (gly-
colide -co-caprolactone) - G-Cap polymer coating on 316 LVM stainless steel in-
creased its corrosion resistance in comparison to non-sterilized ones. Especially in-
crease of breakdown potential and polarization resistance were recorded - table 1.
Moreover it was observed that long-term exposure to the artificial urine decreased
corrosion resistance independently to the sterilization process. Despite NMR and
GCP studies results which shows that the degradation process of the poly(glycolide-
caprolactone) coatings proceeds quite slowly during 90 days, has also been found
that, polymer coating was cracked contributing to unveling of the metal substrate. In
addition especially for 316 LVM samples, it has been observed in the macroscopic
study that prior to exposure, for polymer modified sample, the polymer coating con-
tributes to the retention of corrosion product under its surface and effectively limits
its penetration to the electrolyte. For Ti6Al4V ELI and Ti6Al7Nb alloys with G-Cap
coating, there was no significant effect of radiation sterilization on corrosion resis-
tance parameters determined by potentiodynamic studies. The same as 316 LVM
stainless steel for titanium alloys, long-term exposure to artificial urine resulted in a
decrease of corrosion resistance despite the process of sterilization. In addition, the
polymer coating was removed from both titanium alloys substrate before the end of
the 90-day test period. Separation of the coating from the substrate occurred about
75 days of exposure.
Results of surface mass density of metal ions released to the artificial urine for
all kind of metallic biomaterials, both sterilized and non-sterilized after 90 days of
exposure, showed that sterilization increased the corrosion resistance of the metallic
substrate.
Summarizing the obtained results, it can be stand that biodegradable pol (glycolide-
caprolactone) - G-Cap coating allows for controlled release of the active substance.
Furthermore, radiation sterilization with an electron beam of 25 KGy and an energy
of 10 MeV can be used as a surface treatment of metallic biomaterials intended
for urological implants. Both polymer coating and sterilization process decrease
corrosion resistance of metallic biomaterials substrates. However, it is still neces-
sary to further improve the chemical composition of the G-Cap polymer coating.
Especially proportion of components to ensure better adhesion of the layer to the
metallic substrate during long-term exposure to artificial urine. Also, it is very im-
portant to improve process of polymer coating deposition of metallic substrate to
avoid air bubbles and other disadvantages e.g. uneven layer thickness. It should pro-
vide greater accuracy, repeatability and thickness control.
Acknowledgements
The work has been nanced from research project no BK −201/RIB2/2017(07/020/BK_17/0022).

References
1. Borkowski A.: Urologia: Podr˛ecznik dla studentów medycyny. PZWL, Warszawa (2008)
2. Jacobse S.M., Strickler D.J. Mobley H.L., Shirtliff M.E.: Complicated catheter–assiociated
urinary tract infections due to Escherichia coli and Proteus mirabilis. Clin Microbil Rev 21(1),
26–59 (2008)
3. Shaw G.L., Choong S.K., Fry C.: Ecrustation of biomaterials in the urinary tract. Urol Res
33(1), 17–22 (2005)
4. Ostrowska K., Strzelczyk A., Różalski A., Straczek˛ P.: Biofilm bakteryjny jako przyczyna
zakażeń układu moczowego — mikroorganizmy patogenne, metody prewencji i eradykacji.
Postepy HigMed Dosw 67, 1027–1033 (2013)
5. Klimek L., Różański W., Jabłonowski Z., Sosnowski M., Kliś R.: Obserwacje zmian
mikroskopowych cewników typu Double J w zależności od czasu utrzymywania ich w dro-
gach moczowych. Inzynieria Biomaterialow 43-44, 36–39 (2005)
6. Kajzer A., Kajzer W., Gołombek K., Knol M., Dzielicki J., Walke W.: Corrosion resistance, eis
and wettability of the implants made of 316 LVM steel used in chest deformation treatment.
Archives of Metallurgy and Materials, 61 2a, 767–770 (2016)
7. Szewczenko, J.; Marciniak, J.; Kajzer, W.; Kajzer A.: Evaluation of corrosion resistance of
titanium alloys used for medical implants. Archives of Metallurgy and Materials. 61 2 695–
699 (2016)
8. M. Bartmański, B. Świeczko-Źurek: Project of hip joint endoprosthesis for an individual pa-
tient with materials selection, Advances In Materials Science, 15, 1 (43), 30–36 (2015)
Solid State Phenomena Corrosion and Surface Engineering 227 491–494 (2015)
9. Basiaga M., Paszenda Z., Karasiński P., Kajzer A., Walke W.: Mechanical Properties of An-
odically Oxidized cpTi and Ti-6Al-7Nb Alloy Applications of Computational Tools in Bio-
sciences and Medical Engineering. Adv Struct Mater 71. Chapter 8 Springer 123–132 (2015)
10. Roe D., Karandikar B., Bonn–Savage N., Gibbins B., Roullet J.B.,: Antimicrobial surface
functionalization of plastic catheters by silver nanoparticles. J Antimicrob Chemother 61,
869–876 (2008)
11. Trupti A., Lauserpina C., Shanta M.: Efficacy of a silicone urinary catheter impregnated
with chlorhexidine and triclosan against colonization with proteus mirabilis and other
uropathogens. Infect Control Hosp Epidemiol 28, 596–598 (2007)
12. Hachem R., Reitzel R., Borne A., Jiang Y., Tinkey P., Uthamanthil R., Chandra J., Ghan-
noum M. Raad I.: Novel antiseptic urinary catheters for prevention of urinary tract infections:
correlation of in vivo and in vitro test results. Antimicrob Agents Chemother 53, 5145–5149
(2009)
13. Lee G., Marathe S., Sabbagh S., Crisp J.: Thermo-expandable intra-prostatic stent in the treat-
ment of acute urinary retention in elderly patients with significant comorbidities. Int Urol
Nephrol 37, 501–504 (2005)
14. Opalko F.J., Adria J.H., Khanc S.R.: Heterogeneous nucleation of calcium oxalate trihydrate
in artificial urine by constant composition. J Cryst Growth 181, 410–417 (1997)
15. Szewczenko J., Kajzer W., Grygiel-Pradelok M., Jaworska J., Jelonek K., Nowińska K.,
Gawliczek M., Libera M., Marcinkowski A., Kasperczyk J.: Corrosion resistance of PLGA-
coated biomaterials. Acta of Bioengineering and Biomechanics 19 1 173–179 (2017)
Chapter 42
Mechanical characterization of biodegradable
materials used in surgery
Angela Andrzejewska1
Abstract. Synthetic and natural polymeric materials are widely used in biomedi-
cal and pharmaceutical industries. They are used for manufacturing pharmaceutical
formulations, drug delivery or scaffolds for tissue engineering. The physicochemi-
cal properties of this polymers strongly depend on the percentage of the crystalline
phase. In medical application are used amorphous form of biomaterials. In compar-
ison to a crystalline form of the same material, they are characterized by greater
material properties like: toughness, fatigue strength and hygroscopicity. Due to the
low proportion of the crystalline phase, they have much lower modulus of elasticity,
yield strength, tensile strength and hardness. The application of these materials in
new medical applications should be preceded by gaining knowledge of changing
mechanical properties during hydrolytic degradation. The study used amorphous
form of biodegradable polimer and three degrading solutions: distilled water, Phos-
phate Buffered Saline and sodium chloride 0.9%. The results of the conducted tests
can be used to determine the effect of medium composition on the mechanical prop-
erties of the polymer. Also, they can indicate the optimal degrading medium for
further testing.
Keywords: Amorphous biomaterials, Biodegradable polymers, PLA, Mechanical

properties, Hydrolytic degradation
42.1 Introduction
Available results of the mechanical properties of biodegradable materials are fo-

cused on the study of homogeneous biodegradable polymers and their copolymers,
composites or biodegradable metals. The tested materials are used for the manufac-
turing of sutures [10][14], vascular stents and grafts [2][7] or craniofacial fixation
UTP University of Science and Technology in Bydgoszcz, Al. Prof. S. Kaliskiego 7, 85-796 Byd-
goszcz, POLAND,
e-mail: angela.andrzejewska@utp.edu.pl

https://doi.org/10.1007/978-3-319-70063-2_42
400 Angela Andrzejewska
systems [1][4][13]. In the available studies are also reported the possibilities of using
biodegradable materials as materials in osteosynthesis. Investigation of mechanical
properties are conducted in dry and wet condition for small cross section of the sam-
ples. In terms of experimental studies, the investigation conditions should reflect the
in vivo conditions. The degrading medium which can be use is distilled water [2],
Phosphate Buffered Saline (PBS) [9][10] or simulated body fluid (SBF) [11].
The process of hydrolytic degradation and polymer erosion are associated with
immersion of aliphatic polyester in an aqueous medium. The immersion of speci-
men, causes water absorption by the hygroscopic polymer. The absorbed water pen-
etrates the material and changes the water concentration gradient between the sur-
face and the inner part of the specimen. Result of the interaction of water molecules
and the polymer is the material erosion. The material erosion may occur on the sur-
face or in the entire volume of specimen. This phenomenon is shown in Fig. 42.1.
Homogeneous degradation (bulk erosion) occurs in the material when the rate
of water transport into the inner part of material exceeds the rate of the chemical
changes in the polymer. The phenomenon of homogeneous degradation leads to re-
duction of polymer molecular weight, but the geometrical dimensions of the sample
remain unchanged [5].
Heterogeneous degradation (surface erosion) is associated with a slower rate of
water transport to the inner part of the material in relation to the speed of chemi-
cal changes in the polymer. However, in contrast to the homogeneous degradation,
the heterogeneous degradation leads to dimensional changes of the specimen, but
molecular weight remains unchanged [5].
Fig. 42.1: Bulk and Surface erosion of polymer

42 Biodegradable Materials 401
Concentration gradient between surface and interior part of the material decline
within a few hours or days and its depend on temperature of aqueous medium, spec-
imen thickness, and its hydrophobicity. Small particles like water small particles,
can much faster diffuse than hydrolytic degradation occurs. Therefore, it can be
concluded, that the polymer ester linkages hydrolysis can proceeds evenly [14].
In the case of aliphatic polyesters, the hydrolysis rate is dependent on the am-
bient temperature, molecular structure and material density and also the presence
of ester groups and kind of degrading medium. The degradation rate can be also
dependent on degree of the polymer crystallinity, because water interact primarily
with amorphous domains of polymer [14].
Farrar and Gillson [15] observed that the polyesters hydrolysis, e.g. polylactic
acid, polyglycolic acid and its copolymers, proceed according to the reaction (42.1).
∼ COO ∼ +H2 O COOH ∼ COOH + HO ∼ . (42.1)
Therefore, the first-order kinetics, according to the Michaelis-Menten model is

described in equation (42.2).
d[COOH]
= k [ester][H2 O][COOH] = k[COOH]. (42.2)
dt
The [COOH], [ester], [H2 O] are respectively: the concentration of the carboxylic
acid end groups, esters and water in polimer matrix. It is assumed that, in the initial
phase of the reaction, water and esters concentration are constant [14].
Assuming that [COOH] = 1/Mn, it can be shown that the relationship (42.3) is
proper.
Mn,t = Mn,0 e−kt . (42.3)
The used signs Mn,t and Mn,0 describes the average molecular weight of the poly-
mer at the time “t” and at the beginning of the process. The sign “k” is the average
degree of material hydrolysis [6][14].
It has been shown, that the tensile strength of the polymer can in some cases be
associated with the average molecular weight of the polymer (Mn ). This represents
a equation (42.4).
B B
σ = σ∞ − = σ∞ − . (42.4)
Mnt Mn,0 e−kt
In this case “σ ” describe tensile stress, “σ∞ ” - tensile stress in the infinite molec-
ular weight, B – Empirical Material Constant [6][14].
Within this study was described two main objectives. The first aim of this study
was investigation the mechanical properties of the normative samples with large
cross sections, which are subjected to hydrolytic degradation in three degrading
media. The second objective was to determine the effect of the medium type to
elastic modulus changes, which was designated with σ − ε characteristics.
42.2.1 Material
The study used a biodegradable material Ingeo™ Biopolymer 3100HP (Nature-

Works LLC). Samples with geometry 1A according to the PN-EN ISO 527-1:2012,
was formed by injection molding (Engel e-victory 310/110 injection molding ma-
chine). During sample manufacturing process was adopted parameters, which al-
lowed to obtain an amorphous structure of material.
Specimens groups was immersed in three degrading solutions: distilled water,
Phosphate Buffered Saline and sodium chloride 0.9%, which was heated to 37◦ C.
The pH-value of solutions was monitored to ensure constant value of pH = 7.1. In
each test group minimum five specimens were used. Hydrolytic degradation time
for each group of samples were 1, 6, 12 and 26 weeks.
42.2.2 Mass change
The percentage mass change of samples is calculated from the relationship (42.5),
corresponding to the difference between specimen mass before degradation (m0)
and specimen mass measured after a specified period of degradation (mt ). Sample
weight after degradation is determined for specimens, which was dried with paper
towels in order to remove water from the specimen surface.
mt − m0
Δm = · 100%. (42.5)
m0
42.2.3 Mechanical properties
Static mechanical properties and their changes as a function of degradation time

were determined on the basis of uniaxial tensile test. The testing machine was servo
hydraulic actuator Instron 8874 and strains measurements were taken with the use
of static clamp extensometer. The obtained strain-stress characteristic was used to
determine Young modulus in the linear region of this curve. It is assumed that the
linear region of strain-stress characteristics, can be determined from the ratio of the
values difference expressed in stress (σ1 − σ2 ) and strain points: ε1 = 0.00025 and
ε2 = 0.0005, according to PN-EN ISO 527-1:2012.
42.2.4 Statistical analysis
Statistical analysis was performed using STATISTICA 12.5. To determine the nor-
mality of analyzed results was used the Shapiro-Wilk test, and also was analyzed
the correlation parameter of mass change and the Young’s modulus as a function
of degradation time. In addition, was determined correlation for mass and Young’s
modulus change. The analysis of differences in the measurement results for three
degrading media was performed using ANOVA.
42.3.1 Test of mass change
The average mass changes of the specimens for three degrading media as a function
of degradation time is shown in Fig. 42.2. Noticeably increasing the specimen mass. It
is associated with water absorption by the polymer.
Fig. 42.2: Mass change during the degradation time
In the initial stage of the degradation was observed a much faster water absorp-
tion, than in subsequent degradation stages. It has been demonstrated a linear in-
crease in mass change of samples between successive time points of degradation. A
linear relationship of mass change during the degradation was confirmed by Spear-
man’s rank correlation coefficient. The results of analysis are shown in the table
42.1.
Table 42.1: Correlation of mass change during the degradation time
Spearman’s rank correlation coefficient

Time vs. Mass change
R Spearman t(N-2) p-value
H2 O 0.967 22.34 <0,001
PBS 0.866 10.57 <0,001
NaCl 0.971 24.77 <0,001
In 1st , 6th and 12th degradation week, were no statistically significant differences
in specimenâĂŹs mass change immersed in various media degradation. Statistically
significant differences in specimenâĂŹs mass changes were observed in 26th degra-
dation week. The highest rate of mass change in T26 are characterized specimens
immersed in distilled water, while the smallest percentage mass change character-
izes by specimens immersed in Phosphate Buffered Saline (PBS). The results are
shown in the table 42.2.
Table 42.2: Results of descriptive statistics and ANOVA, specimens mass change,
%
Week Degrading Average Standard Minimum Median Maximum p - value

medium Value Deviation Value Value
H2 O 0.00 0.00 0.00 0.00 0.00
T0 PBS 0.00 0.00 0.00 0.00 0.00 -
NaCl 0.00 0.00 0.00 0.00 0.00
H2 O 0.01 0.00 0.01 0.01 0.01
T1 PBS 0.01 0.00 0.01 0.01 0.01 0.108
NaCl 0.01 0.00 0.01 0.01 0.01
H2 O 0.01 0.00 0.01 0.01 0.01
T6 PBS 0.01 0.00 0.01 0.01 0.01 0.589
NaCl 0.01 0.00 0.01 0.01 0.01
H2 O 0.01 0.00 0.01 0.01 0.01
T12 PBS 0.01 0.00 0.01 0.01 0.01 0.105
NaCl 0.01 0.00 0.01 0.01 0.01
H2 O 0.02 0.00 0.01 0.01 0.02
T26 PBS 0.01 0.00 0.01 0.01 0.01 0.002
NaCl 0.01 0.00 0.01 0.01 0.02
42.3.2 Test of static properties
The average Young modulus changes of the specimens for the three degrading media
as a function of degradation time is shown in Fig. 42.3.
Fig. 42.3: The average Young modulus changes
In the initial stage of the degradation âĂŞ to the 6th week, was observed a de-
crease in the Young modulus. But during the longer degradation times (12th and
26th week) exhibits its growth. The results are characterized by high variability and
relatively low value of the correlation coefficient. It is described by the data in table
42.3.
Table 42.3: Correlation of mass change during the degradation time
Spearman’s rank correlation coefcient

Time vs. Mass change
R Spearman t(N-2) p-value
H2 O 0.303 1.883 0.068
PBS 0.370 2.425 0.020
NaCl 0.449 3.057 0.004
Comparison of the results differences of Young’s modulus for specimens im-

mersed in the degrading media, showed the constancy of this parameter in the 1st ,
6th and 26th week. By contrast to the earlier results, in the T12 was observed a higher
value of Young’s modulus for specimens immersed in 0.9% NaCl solution than in
the case of specimens immersed in distilled water or PBS. The results are presented
in table 42.4.
Table 42.4: Results of descriptive statistics and ANOVA, specimens mass change,
%
Week Degrading Average Standard Minimum Median Maximum p - value

medium Value Deviation Value Value
H2 O 3508.9 109.4 3353.5 3474.2 3707.0
T0 PBS 3508.9 109.4 3353.5 3474.2 3707.0 -
NaCl 3508.9 109.4 3353.5 3474.2 3707.0
H2 O 3460.6 133.2 3291.0 3431.3 3653.8
T1 PBS 3495.8 114.9 3332.3 3463.8 3690.8 0.801
NaCl 3506.5 141.9 3380.5 3442.7 3770.3
H2 O 3415.3 45.4 3345.6 3425.4 3484.8
T6 PBS 3399.5 49.2 3306.3 3406.9 3450.2 0.369
NaCl 3439.6 70.0 3332.1 3432.0 3572.1
H2 O 3528.5 80.5 3456.5 3504.7 3690.9
T12 PBS 3524.6 44.9 3456.4 3530.1 3593.7 0.025
NaCl 3599.3 32.8 3538.1 3603.3 3640.3
H2 O 3638.1 206.0 3451.0 3614.0 4082.6
T26 PBS 3699.8 230.0 3457.5 3621.5 4202.6 0.827
NaCl 3676.6 161.6 3510.0 3598.2 3877.5
42.3.3 Comparison between mass change and mechanical

properties
It has been observed, that absorbed degrading solution causes changes in mechanical
parameters of the tested material. The effects of fluid absorption and degradation
phenomenon, may be the formation of microcracks in the material. It is believed,
that the crack nucleation is induced in water-soluble inclusions of material and its
promote with increasing osmotic pressure. The osmotic cracking phenomenon was
described in [3][6][10][12].
The σ -ε characteristic designated for samples incubated in various degrading
solutions allows characterize of the testing material. The specimens, which were
subjected to uniaxial stretching before hydrolytic degradation, exhibit σ -ε charac-
teristics corresponding to the characteristics of brittle materials. An when specimens
were degraded, the σ -ε characteristics corresponding to characteristic of polymers
with the yield point. The change of σ -ε characteristic for the degrading specimens
is related to the plasticizing action of the degrading solution.
The different polymer behavior immersed in the physiological fluids environment
(PBS and 0.9% NaCl) and transient changes in the physicochemical and mechanical
properties, in relation to the specimens immersed in distilled water may be affected
by the phenomenon of crystallization, occurring due to the arrangement of bonding

the polymer. And also the presence of additional substances, i.e., Sodium chloride
and phosphoric acid salts (in buffered saline).
42.4 Conclusions
The available literature, presents results of the biodegradable materials degradation,

including the results of degradation of polylactide [1]. In the present study, authors
are focus on the determination of the effect of medium composition and ratio of its
absorption on a change of Young’s modulus of biodegradable polymer specimens in
the monomer form and having a larger cross section than test specimens presented
in the literature.
It is assumed that the obtained results may find a practical use in determining the
suitability of the tested materials in medical applications other than, e.g. Surgical
sutures, vascular stents or microplates cranial and maxilofacialsurgery.
The determining of the effect of the medium composition on material mechanical
properties changes, becomes important from the point of view of planning experi-
ments. In the literature are presented results for degradation with use of one type of
degrading medium, which is used interchangeably from the three indicated herein. It
is noted that the presence of salt ions in the degrading medium may affect the differ-
ent material behavior, in relation to the medium devoid of additional ions. For this
reason, the optimal choice of medium simulating the in vivo conditions of the im-
plant works, would be selected the degradation solution similar to the composition
of body fluids.
The determined σ -ε characteristics can also find practical application in the im-
mediate definition of material model during modeling biodegradable implants and
their behavior by the degradation time with using finite element methods.
References
1. Andrzejewska A., Topoliński T.: Polimery biodegradowalne do zastosowań biomedycznych.

Post˛epy w inżynierii mechanicznej. 6(3), 5–12 (2015)
2. Bartkowiak-Jowsa M., B˛edziński R., Kozłowska A., Filipiak J., Pezowicz C.: Mechanical, rhe-
ological, fatigue, and degradation behavior of PLLA, PGLA and PDGLA as materials for vas-
cular implants. Meccanica. 48, 721–731 (2013)
3. Bélan F., Belleger V., Mortaigne B.: Hydrolytic stability of unsaturated polyester networks with
controlled chain ends. Polymer Degradation and Stability. 56, 93–102 (1997)
4. Buijs G.J., Houwen E.B., Stegenga B., Verkere G.J., Bos R.R.M: Mechanical Strenght and
Stiffness of the Biodegradable Sonic Weld Rx Osteofixation System. Journal of Oral and Max-
illofacial Surgery. 67, 782–787 (2009)
5. Dłuska E., Markowska-Radomska A.: Analiza mechanizmów uwalniania składnika aktywnego
z utrwalonych do postaci mikrosfer emulsji wielokrotnych. Inżynieria i Aparatura Chemiczna.
49(1), 33–34 (2010)
6. Farrar D.F., Gilson R.K.: Hydrolytic degradation of polyglyconate B: the relationship between
degradation time, strength and molecular weight. Biomaterials. 23, 3905–3912 (2002)
7. Gautier L., Mortaigne B., Bellenger V., Verdu J.: Osmotic cracking nucleation in hydrothermal-
aged polyester matrix. Polymer. 41, 2481–90 (2000)
8. Milewski K., Tajstra M.: Stenty bioresorbowalne âĂŤ aktualny stan wiedzy. Folia Cardiologica
Excerpta. 7, 213–219 (2012)
9. Olewnik-Kruszkowska E.: Influence of the type of buffer solution on thermal and structural
properties of polylactide-based composites. Polymer Degradation and Stability. 129, 87–95
(2016)
10. Pantani R., Sorrentino A.: Influence of crystallinity on the biodegradation rate of injection-
moulded poly(lactic acid) samples in controlled composting conditions. Polymer Degradation
and Stability. 98, 1089–1096 (2013)
11. Roethera J.A., Boccaccinib A.R., Henchb L.L., Maquetc V., Gautierc S., Jerome R.: Devel-
opment and in vitro characterisation of novel bioresorbable and Bioactive composite materials
based on polylactide foams and Bioglasss for tissue engineering applications. Biomaterials. 23,
3871–3878 (2002)
12. Saylor D.M., Richardson D.C., Dair B.J., Pollack S.K.: Osmotic cavitation of elastomeric
intraocular lenses. Acta Biomaterialia. 6, 1090-âĂŞ1098 (2010)
13. Väänänen P., Nurmi J.T., Lappalainen R., Jank S.: Fixation properties of a biodegradable “free-
form” osteosynthesis plate with screws with cut-off screw heads: Biomechanical evaluation
over 26 weeks. Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology. 107, 462–
468 (2009)
14. Vieira A.C., Guedes R.M., Tita V.: Constitutive modeling of biodegradable polymers: Hy-
drolytic degradation and time-dependent behavior. International Journal of Solid and Structures.
51, 1164–1174 (2014)
15. Vieira A.C., Vieira J.C., Guedes R.M., Marques A.T.: Degradation and Viscoelastic properties
of PLA-PCL, PGA-PCL, PDO and PGA fibers. Materials Science Forum. 636-37(1), 825–832,
(2010)
Author Index
A Gacek, Adam, 117, 151

Adamowicz-Czoch, Elżbieta, 117 Gadula-Gacek, Elżbieta, 117
Andrzejewska, Angela, 399 Gałecka, Jerzy, 151
Antonowicz, Magdalena, 329, 359 Galińska, Marta, 23, 33
Arkusz, Katarzyna, 339 Garbacik, Małgorzata, 193
Gąsior, Mariusz, 117
B Głąb, Ewelina, 311
Badura, Paweł, 3 Glowiński, Sebastian, 235
Bartel, Sebastian, 141 Górka, Wioletta, 217
Basiaga, Marcin, 329 Grabski, Jakub Krzysztof, 53
Bednarek, Ilona, 13 Grochala, Dominik, 205
Bieniek, Andrzej, 245, 299 Gwóżdź, Arkadiusz, 117
Blażejewski, Andrzej, 235 Gzik, Marek, 245, 271
Bodusz, Wojciech, 3 Gzik-Zroska, Bożena, 271
Bogacz, Wojciech, 359
Budzianowski, Zbigniew, 193 H
Bugdol, Marcin D., 107 Haq, Anam, 43
Bugdol, Monika N., 107
J
C Jamrozik, Wojciech, 319
Chrzan, Miłosz, 245, 299 Jaworska, Joanna, 391
Cieślak, Magdalena, 53 Jędrusik, Przemysław, 13
Czajkowska, Joanna, 23, 33 Jelonek, Katarzyna, 391
Jochymczyk-Woźniak, Katarzyna, 217, 225
D Joszko, Kamil, 245, 271
Dąbrowski, Jan Ryszard, 369 Juszczyk, Jan, 23, 33, 87
Derejczyk, Jarosław, 245
Długosz, Magdalena, 245 K
Drewniak, Sabina, 359 Kajor, Marcin, 205
Kajzer, Anita, 311, 391
F Kajzer, Wojciech, 311, 391
Fabian, Emil, 61 Kalarus, Zbigniew, 117
Filipowska, Anna, 171 Kamieniczna, Marzena, 61
Filipowski, Wojciech, 171 Kántoch, Eliasz, 205
Flak, Tomasz, 161 Kaspera, Wojciech, 271
Fritzkowski, Paweł, 263 Kawlewska, Edyta, 271
Kiel-Jamrozik, Marta, 319
G Komorowski, Dariusz, 125
Gabor, Jadwiga, 161 Konopelska, Agnieszka, 225

https://doi.org/10.1007/978-3-319-70063-2
410 Author Index
Koprowski, Robert, 13 Poloński, Lech, 117

Kostera, Kinga, 79 Prochaczek, Fryderyk, 151
Kostka, Paweł, 133 Pyciński, Bartłomiej, 87
Kowalski, Pawel, 69
Krawczyk, Cezary, 359 R
Kręcichwost, Michał, 79 Raif, Paweł, 185
Krżyzyński, Tomasz, 235 Ranosz, Żaneta, 33
Kubik, Klaudia, 161 Roj, Dawid, 69
Kucharski, Dariusz, 205 Rozentryt, Piotr, 117
Kucharski, Dawid, 97
Kurpisz, Maciej, 61 S
Sambok, Agata, 359
L Semenowicz, Jacek, 311
Lackner, Jürgen Markus, 379 Sidun, Jarosław, 369
Łagan, Sylwia, 255, 279 Sieciński, Szymon, 133
Łężniak, Marta, 161 Smreczak, Tomasz, 107
Liber-Kneć, Aneta, 255, 279 Sobkowiak, Marta, 271
Luszawski, Jerzy, 225 Sobotnicki, Aleksander, 69
Sokal, Adam, 117
M Stachowska, Ewa, 61
Majewska, Justyna, 359 Stopa, Marcin, 263
Major, Bogusław, 379 Swinarew, Andrzej S., 161
Major, Roman, 379 Swinarew, Beata, 161
Mandera, Marek, 225 Szczurek, Zbigniew, 69
Michałowska, Martyna, 53, 263 Szewczenko, Janusz, 391
Michnik, Robert, 217, 225, 299 Szuster, Barbara, 69
Mika, Barbara, 125, 161
Miodońska, Zuzanna, 3 T
Mitas, Andrzej, 299 Tajstra, Mateusz, 117
Moczulski, Wojciech, 289 Tkacz, Ewaryst, 171, 185, 193
Muzalewska, Małgorzata, 289 Trawiński, Tomasz, 141
Trembecka-Wojciga, Klaudia, 379
N
Nakonieczny, Damian S., 359 W
Nawrat-Szołtysik, Agnieszka, 217 Walczak, Tomasz, 53, 263
Niedziela, Jacek, 117 Walke, Witold, 329
Nowak, Jagoda, 97 Wi˛eclawek, Wojciech, 79
Nowakowska, Katarzyna, 217, 225, 299 Wijata, Agata, 23, 33
Nowińska, Katarzyna, 391 Wilczek, Marcin, 117
Wilk, Szymon, 43
O Winiarska-Prochaczek, Hanna, 151
Obuchowicz, Rafał, 87 Witkowska, Ilona, 319
Ociessa, Aneta, 117 Wodarski, Piotr, 245, 299
Ogiegło, Weronika, 23 Wojciechowski, Konrad, 117
Okła, Hubert, 161 Wolański, Wojciech, 271
Oleksy, Wojciech, 193 Woloszyn, Jakub, 69
Woźniak, Anna, 349
P Wróbel, Tomasz, 311
Paluch, Jarosław, 161 Wróbel, Zygmunt, 13
Paszenda, Zbigniew, 329 Wujec, Monika, 171
Paszenda, Zbigniew K., 359
Pilch, Zbigniew, 141 Z
Płaczek, Aleksander, 117 Ziębowicz, Anna, 349
Pollak, Anita, 299 Ziębowicz, Bogusław, 329, 349

(Advances in Intelligent Systems and Computing 623) Marek Gzik, Ewaryst Tkacz, Zbigniew Paszenda, Ewa Piętka (Eds.) - Innovations in Biomedical Engineering-Springer International Publishing (2018)

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

(Advances in Intelligent Systems and Computing 623) Marek Gzik, Ewaryst Tkacz, Zbigniew Paszenda, Ewa Piętka (Eds.) - Innovations in Biomedical Engineering-Springer International Publishing (2018)

Diunggah oleh

Hak Cipta:

Format Tersedia

Advances in Intelligent Systems and Computing 623

More information about this series at http://www.springer.com/series/11156

Zbigniew Paszenda Ewa Piętka

ISSN 2194-5357 ISSN 2194-5365 (electronic)

© Springer International Publishing AG 2018

Printed on acid-free paper

This Springer imprint is published by Springer Nature

success in implementing consequences of human imagination into the social life. We

Zabrze, Ewaryst TKACZ

Part I Informatics in medicine

1 Approach for spectrogram analysis in detection of selected

3 Breast Cancer Segmentation Method

7 Human Sperm Morphology Analysis using a Digital Holographic

9 Prototype measurement system for spatial analysis of speech signal

10 Shear Wave Elastography and Strain Elastography: a Study on a

Part II Signal analysis

13 2017 Monitoring and Teletransmission of Medical-Data in Heart

13.2.2 EXTRABIOTEL-HF study . . . . . . . . . . . . . . . . . . . . . . . . . 120

15 Determining Heart Rate Beat-to-Beat from Smartphone

18 Novel tumor protein markers collection by the use of highly porous

21.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

Part III Modelling and simulations in biomechanics

23 Assessment of balance of older people living at a social welfare home 217

25 Body part accelerations evaluation for chosen techniques in martial

28 Flow of subretinal liquid through the retinal hole after surgery –

29 Impact of Vessel Mechanical Properties on Hemodynamic

Part IV Engineering of biomaterials

33 Assessment Of Hip Endoprosthesis Cups After Reimplantation . . . . . 311

35 The Funcionalization of Grade 4 Surface Used for Blood Contacting

36 The inﬂuence of implantation on mechanical degradation of the

37 The Inﬂuence of Technology on the Physicochemical and

38.2.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

41 Efect Of Sterylization And Long-Term Exposure To Artiﬁcial

42.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409

Katowice - Szpital Geriatryczny im. Jana Pawła II

Centre of Polymer and Carbon Materials of the Polish Academy of Sciences

Department of Biomechatronics, Faculty of Biomedical Engineering,

30 Mickiewicza Av., 30-059 Kraków, Poland

Medyków 16, 40-751 Katowice, Poland

Institute of Psychology, University of Silesia, Katowice, Poland

Silesian University of Technology, ul. F. D. Roosevelta 40, 41-800 Zabrze, Poland,

Wojciech Bodusz1 , Zuzanna Miodońska1 , and Paweł Badura1

Abstract. An attempt to automatise selected pronunciation pathology detection in

Keywords: speech pathology, spectrogram analysis, image processing

Faculty of Biomedical Engineering, Silesian University of Technology,

© Springer International Publishing AG 2018 3

cessing methods, such as filtration, thresholding or edge detection algorithms [9].

1.2 Materials and methods

Fig. 1.1: General processing workow.

from in Figs. 2(a) and 2(b), respectively.

(a) (b) (c)

1.2.2 Image processing and extraction of image features

Further analysis involves processing of spectrogram as a monochromatic image.

1.2.2.1 Image centroid

1.2.2.2 Average centroid crossing ratio (ACCR)

1.2.2.4 Average gray level

1.2.2.5 Standard deviation

1.3 Results and discussion

1.3.1 Experimental results

Table 1.1: Algorithm performance in the 4-fold cross-validation.

Metrics Accuracy Sensitivity Speci f icity

Table 1.2: Algorithm performance in the leave-one-out cross-validation.