Teaching  

The Ohio State University

 
line decor
  HOME  ::  TEACHING  :: CBCSL
line decor
   

 

Bayes Optimality in Pattern Recognition

The handouts and papers you will find here are for the tutorials “Feature Extraction and Classification” (at IEEE CVPR 2007 conference) and “Bayes Optimality in Statistical Pattern Recognition” (at IbPRIA 2007). The handouts are an extension of the talks given from 2005 to 2007 at different Universities under the title “The Secret Life of Linear Methods: Why they work, do not work, and can be made to work.”

Download:

  • Handouts.
  • Audio format (will be available later in the year).

Please note: These handouts are for personal use only. Some of the images and figures used here are copyrighted. If you want to use these, you will need to get permission from the appropriate place.

This tutorial is based on the following papers:

  1. Martinez & Kak, “PCA versus LDA,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 23, No. 2, pp. 228-233, 2001.
  1. Martinez & Zhu, “Where are linear feature extraction methods applicable,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 27, No. 12, pp. 1934-1944, 2005.
  1. Zhu & Martinez, “Subclass Discriminant Analysis,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 28, No. 8, pp. 1274-1286, 2006.
  1. Zhu & Martinez, “Pruning Noisy Bases in Discriminant Analysis,” IEEE Transactions Neural Networks, Vol. 19, No. 1, pp. 148-157, 2008.
  1. Hamsici & Martinez, “Bayes Optimality in Linear Discriminant Analysis,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 30, No. 4, pp. 647-657, 2008.
  1. Hamsici & Martinez, “Spherical-Homoscedastic Distributions: The equivalency of spherical and Normal Distributions in classification,” Journal of Machine Learning Research, 8(Jul):1583-1623, 2007.

If you have made improvements over these results please let me know and I will add your results in subsequent seminars. General comments are also welcomed. Email them to me.

Acknowledgments: This research was partially supported by the National Institutes of Health and the National Science Foundation.