Computer Vision - Spring 2006
Lecture 1: Support Vector Machines and Plankton Classification (April 25, 2006)
- Email me for a hardcopy of the SVM lecture notes.
- Here is a set of SVM tutorials and slides worth checking out:
- "A Tutorial on Support Vector Machines for Pattern Recognition", by Chris Burges
- "A Short Introduction to Learning with Kernels", Bernhard Scholkopf and Alex Smola
- A tutorial on "SVM's and Kernel Machines" by Nello Cristianini presented at ICML 2001
- The Support Vector website
- The Kernel Machines website
- Google!
- Slides (.ppt) for the second half of the lecture.
- Mini-homework due by the end of the week.
- You can browse the Caltech-101 data set here, instead of downloading it!
- Try running this Matlab function. This is a view of figure 2 (read the comments in the Matlab file).
- Also, try playing around with this awesome SVM applet.
- PCA notes are forthcoming.
- Check out the PCA entry on Wikipedia.
- Gilbert Strang's linear algebra video lectures are incredible.
- Chris Burges has an awesome tutorial on important mathematical concepts including lagrange multipliers.
- Homework assigned (due 05/16 in-class).
- Helper Matlab function for Problem 2 and sample
output.
- Link to Eigenfaces journal paper.
- This website has a list of video lectures on various topics in machine learning, including ICA. Definitely worth taking a look at.
- Check out these ICA tutorials and slides:
- Matlab code for RADICAL: An ICA algorithm developed by Erik Learned-Miller and John Fisher
- Here are links to the papers I mentioned in class:
- Learning Depth from Single Monocular Images,
Ashutosh Saxena, Sung Chung, and Andrew Y. Ng.
In Neural Information Processing System 2005 - Recognizing Panoramas,
Matthew Brown and David Lowe
In International Conference on Computer Vision 2003
Also, check out Matthew Brown's website for a downloadable demo. - EM Algorithms for PCA and SPCA,
Sam Roweis
In Neural Information Processing Systems 1997 - Face Recognition in Subspaces,
G. Shakhnarovich and B. Moghaddam
Handbook of Face Recognition, 2004
- Learning Depth from Single Monocular Images,
- Matlab resources:
- Matlab Central - code repository
- Matlab blog by Matlab language designer Loren Shure
- Tom Minka's notes on accelerating Matlab. Pay particular attention to using the profiler and writing vectorized code.
- Marios Athineos's tips and tricks.
- comp.soft-sys.matlab Google group is a great resource.
- I recommend running Matlab in command-line mode (by typing "matlab -nojvm -nosplash") and using emacs to edit your M-files. This way you will eat up less resources. Also, download matlab.el to get Matlab style editing in emacs. Read the comments to see how to edit your .emacs file. Here is my .emacs file.
- The course website has a section on Matlab tutorials, check it out!
- Email me if you have any Matlab questions.
BEST OF LUCK TO ALL GRADUATING SENIORS!!!!!!
