Semaine 1 (Week 1) dimanche

~ dimanche ~

• Segmentation
• Wrote shell script to batch run segmentation with JSEG. While there are over-segmented areas, the region between each ingredient boundary is clearly segmented, which is good. However, the plate is not that clearly separated from the food, as the plate is often also over-segmented, so we'll need a way to discard the plate and tabletop.

• Brightness
• Tune all images to the same brightness, might be able to deal with variations in lighting this way?
• Result: Segmentation is better WITHOUT brightness adjustment, but don't know about texton extraction yet.
• Motivation for brightness adjustment is intra-class variation in lighting. Color textons would use color information and misclassify if there's too much variation. Ting-Fan's cafeteria vision ended up depending heavily on the color, so that's very important. Will see if need this brightness tuning when texton is running.
• TODO: Attach images 

• Textons
• "basic elements in early (pre-attentive) visual perception" [3]
• [3]'s process basically goes from input image -> generate textons -> generate base map -> generate an image, from a combination of bases, that tries to match the input image
• Everyone seems to be using k-means
• What are Textons? http://chengenguo.com/ucla/what_are_textons.htm
• paper http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.76.8748&rep=rep1&type=pdf [3]
• Varma, Zisserman paper http://research.microsoft.com/en-us/um/people/manik/pubs/varma05.pdf [4]
• filter bank code, MATLAB http://www.robots.ox.ac.uk/~vgg/research/texclass/filters.html
• Texton Code in MATLAB from Ting-Fan
• Textonizer found on Google Code. C++, OpenCV 1.0
• Textonator.h: http://code.google.com/p/texton-izer/source/browse/trunk/src/Textonator.h?spec=svn66&r=66
• Textonator.cpp: http://code.google.com/p/texton-izer/source/browse/trunk/src/Textonator.cpp?r=66
• main.cpp: http://code.google.com/p/texton-izer/source/browse/trunk/src/main.cpp?spec=svn38&r=29
• Berkeley Segmentation Dataset and Benchmark - Textons > computeTextons.m http://www.eecs.berkeley.edu/Research/Projects/CS/vision/bsds/
• Texture Synthesizer, C++, OpenCV http://texsynth.googlecode.com/svn-history/r3/trunk/TextureSynthesizer/TextureSynthesizer.cpp 
• Deciding on which to use, what the differences are, and which filter bank to use. Ting-Fan seems to have used the John Winn filter bank, there's also the Leung-Malik filter bank and the Schmid filter bank. [4] used Leung-Malik. Don't know what the differences are yet. Guess will just try one and see how (bad) it goes first for starters to save time.
• Others
• Toolbox for Computer Vision http://gulimujyujyu.me/wiki/index.php?title=Toolbox_for_Computer_Vision

[3] Song-Chun Zhu, Cheng-En Guo, Yizhou Wang, and Zijian Xu. What are Textons? IJCV 2005. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.76.8748&rep=rep1&type=pdf

[4] Manik Varma and Andrew Zisserman. A Statistical Approach to Texture Classification from Single Images. http://research.microsoft.com/en-us/um/people/manik/pubs/varma05.pdf