Generating PowerPoint Slides from Hand-Drawn Sketches

Jeff Wheeler - jeffwheeler@stanford.edu

Muneeb Ahmed - ahmedm@stanford.edu

Description
As engineers, we often find ourselves spending a non-trivial amount of time converting sketched
diagrams and figures into digital format. To solve this problem, we aim to convert hand-drawn
sketches (including shapes, texts, arrows, etc.) to PowerPoint slides. The goal is to accomplish
this by using techniques such as OCR and shape recognition algorithms. Upon searching the
internet, we came across several papers on shape-recognition but none of them are in the
context of presentation slides. There are several algorithms out there that aim to detect shapes
and other hand-drawn figures. We aim to bring together these algorithms along with our own to
make an effective program for conversion of hand-drawn sketches to PowerPoint presentations.

Implementation
We intend to initially detect slide boundaries and then detect the interior contents after aligning
the frame based on the boundaries. We will be able to detect shapes of variable size via the
linear components and a collection of other similar algorithms.
We intend to capture photos using Android, then send photos to a server to generate slides and
combine them into a PPT presentation, and return them to the phone for viewing and sharing
(e.g. via email). The server side code will perform all image recognition and formatting for the
resulting presentation.

References
Jakowski, Wojciech and Krawiec, Krzysztof and Wieloch, Bartosz. 2007. "Learning and
Recognition of Hand-Drawn Shapes Using Generative Genetic Programming." In
Applications of Evolutionary Computing, 281-290. Springer Berlin Heidelberg.
Levent Burak Kara, Thomas F. Stahovich. 2005. "An image-based, trainable symbol recognizer
for hand-drawn sketches." Computers & Graphics, August: 501-517.
Seong-Whan Lee, Jin H. Kim, Frans C. A. Groen. 1990. "Translation-, Rotation- And ScaleInvariant Recognition Of Hand-Drawn Symbols In Schematic Diagrams." International
Journal of Pattern Recognition and Artificial Intelligence.
Yan Ping Zhou, Chew LIm Tan. 2000. "Hough Technique for Bar Charts Detection and
Recognition in Document Images." International Conference on Image Processing.
Vancouver. 605-608.