SIDDAGANGA INSTITUTE OF TECHNOLOGY

DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING

PRESENTATION ON SPIHT ALGORITHM

BY

NEERAJ KUMAR (1SI09EC061)

UNDER THE GUIDANCE OF

SWETHA N. , M.Tech.,

CONTENTS

INTRODUCTION OBJECTIVE IMAGE COMPRESSION WAVELET TRANSFROM WAVELET DECOMPOSITION SPIHT CODEC FLOW CHART NUMERICAL RESULTS APPLICATIONS CONCLUSION

INTRODUCTION

It is a fast and efficient method with good image quality,

high PSNR, especially for color images.

Produces a fully embedded coded file. Simple quantization algorithm. Fast coding/decoding algorithm. It can be used for lossless compression. It can code to exact bit rate or distortion.

OBJECTIVE

Digital information must be stored, retrieved, analyzed

and processed in an efficient manner, in order for it to be put to practical use.

The bandwidth required to transmit the image of size

720*1280 pixels is very large so we need to compress these images in order to transmit them without wasting the bandwidth.

IMAGE COMPRESSION

Image compression is technique under image processing

having wide variety of applications.

The fundamental components of compression are redundancy and irrelevancy reduction.

Redundancy means duplication.

Irrelevancy means the parts of signal that will not be noticed by the Human Visual System. Image compression focuses on reducing the number of bits needed to represent an image.

WAVELET TRANSFORM

It is used to provide multiresolution analysis. The DWT analyzes the signal at different frequency bands with different resolutions by decomposing the signal into a coarse approximation and detail

information.

It employs two sets of functions, called scaling functions

and wavelet functions, which are associated with low

pass and high pass filters.

Fig 3:-DWT coefficients at different levels

WAVELET DECOMPOSITION

The level of decomposition is given by: - level=log2n, n is the number of pixels in a given row or column.

It produce a pyramid structure where an image is

decomposed sequentially by applying low pass and high pass filters and then decimating the resulting images.

These are one-dimensional filters that are applied in cascade (row then column) to an image.

It creates a four-way decomposition: LL, LH, HL and finally HH The resulting LL version is again four-way

decomposed. This process is repeated until the top of the

pyramid is reached.

Fig 4:- Image decomposition using wavelets

SPIHT CODEC

There exists a spatial relationship among the coefficients

at different levels in the pyramid structure.

A wavelet coefficient at location (i,j) in the pyramid representation has four direct descendants (off-springs) at locations: O(i,j)={(2i,2j),(2i,2j+1),(2i+1,2j),(2i+1,2j+1)}

This pyramid structure is commonly known as spatial

orientation tree.

Fig 5:-: Off-spring dependencies in the pyramid structure

ENCODING/DECODING ALGORTIHM

O(i,j): set of coordinates of all offspring of node (i,j);

children only

D (i,j): set of coordinates of all descendants of node (i,j); children, grandchildren, great-grand, etc.

H (i,j): set of all tree roots (nodes in the highest pyramid level); parents

L (i,j): D (i,j) O(i,j) (all descendents except the

offspring); grandchildren, great-grand, etc.

Initialization:-

n = log2 (max |coeff|)

LIP = All elements in H LSP = Empty LIS = Ds of Roots Step 1: Initialization: Set n to target bit rate.

for each node in LIP do:

if Sn [ i, j] = 1, move pixel coordinates to the LSP and keep the sign of c(i,j) ;

Significance Map Encoding (Sorting Pass) Process LIP for each coeff (i,j) in LIP Output Sn(i,j) If Sn(i,j)=1, Output sign of coeff(i,j): 0/1 = -/+ Move (i,j) to the LSP

End if
End loop over LIP Process LIS for each set (i,j) in LIS if type D Send Sn(D(i,j))

If Sn(D(i,j))=1 for each (k,l) O(i,j), output Sn(k,l) if Sn(k,l)=1, then add (k,l) to the LSP and output sign of coeff: 0/1 = -/+ if Sn(k,l)=0, then add (k,l) to the end of the LIP

end for
End if else (type L ) Send Sn(L(i,j)) If Sn(L(i,j))=1 add each (k,l) O(i,j) to the end of the LIS as an entry of type D

remove (i,j) from the LIS end if on type End loop over LIS Refinement Pass Process LSP for each element (i,j) in LSP except those just added above Output the nth most significant bit of coeff End loop over LSP

Update
Decrement n by 1 Go to Significance Map Encoding Step
..\kedia\SPIHT_Charts.pdf

FLOW CHART

NUMERICAL RESULTS

PSNR: P1=34.7835 P2=35.2734 P3=34.5560 TOTAL=34.8710

Fig 6: Compression of Lena color image with rate=1

PSNR: db1-35.05 db4-35.04 bior4.4-35.56 sym2-34.93

Fig 7: Compression of cameraman image with rate=1

Comparison of different images

RATE/ IMAGES Lena.jpeg
Sunset.jpeg Fruits.jpeg Tulips.jpeg

1.0 bpp 34.8710

39.1100 34.2636 33.4383

0.75bpp 32.9807
37.7216 32.5162 31.1407

0.5bpp 30.3955
35.7141 30.1835 28.7387

0.25bpp 27.2055
32.8820 27.0384 25.3465

Results on Lena image

COMPONENT / RATE 111 Y Cb Cr

34.8735

35.2734

34.5560

0.5 0.5 0.5

1 0.5 0.5 1 0.2 0.2 0.5 1 1 0.2 1 1

30.3266
34.6729 34.2094 30.3674 26.5395

30.4771
35.1478 34.9236 30.5519 26.5875

30.3827
34.4722 33.5824 30.2674 26.4919

Results on grayscale image

RATE/ FILTERS
db1

1.0 bpp

0.75bpp

0.5 bpp

0.25bpp

35.05

32.26

29.68

26.46

coif1

35.07

32.44

29.94

26.50

sym2

34.93

32.28

29.69

26.39

bior4.4

35.56

33.01

30.50

27.13

COMPARISON OF EZW & SPIHT

RATE COMPRESSION EZW SPIHT

1.0
0.5 0.25 0.125 0.0625

8:1
16:1 32:1 64:1 128:1

39.55
36.28 33.17 30.23 27.54

39.92
36.68 33.38 30.40 27.69

APPLICATIONS

SPIHT has been successfully tested in natural (portraits,

landscape, weddings, etc.) and medical (X-ray, CT, etc)

images.

It is effective in a broad range of reconstruction qualities. It

can code fair-quality portraits and high-quality medical

images equally well.

It is used in compression of elevation maps, scientific data. It is also being used in case of ECG signals.

CONCLUSION

SPHIT algorithm uses the principle of partial ordering by

magnitude, set partitioning by significance of magnitudes

with respect to a sequence of octavely decreasing threshold, ordered bit-plane transmission, and self-similarity across

scale in an image wavelet transform.

The realization of these principles in matched encoding and decoding algorithms is a new one and is shown to be effective than in previous implementations of EZW algorithm.

REFERENCE

AMIR SAID AND WILLIAM A. PEARLMAN, A New, Fast, and Efficient Image Codec Based on Set Partitioning in Hierarchical Trees. IEEE Transaction on circuits & systems for video technology Vol. 6 No. 3, 1996.

J. M. Shapiro, Embedded image coding using zerotrees of wavelet coefficients. IEEE Trans. Signal Processing vol.41 pp 3445-3462, Dec 1993.

ALDO MORALES AND SEDIG AGILI, Implementing the SPIHT Algorithm in MATLAB. Proceedings of the 2003 ASEE/WFEO International Colloquium

J. MAL, P. RAJMIC,DWT-SPIHT image codec implementation.

JAMES

S.

WALKER,

Wavelet-based

Image

Compression.

KAHLID SAYOOD, SPIHT_CHARTS.

ROBI POLIKAR,Wavelet tutorial.

WAVELET TRANSFORMS by Raghuveer M Rao.

FUNDEMENTALS OF MULTIMEDIA by Ze-Nian

Li and Mark S Drew.

THANK YOU

ANY QUERRIES???