Physics Colloquium, November 4, 2003

Image Analysis and Other Adventures Using Empirical Eigenfunctions (SVD, PCA, KL, etc.)

Lawrence Sirovich, Professor
Director, Laboratory of Applied Mathematics
Department of Biomathematics

Mt. Sinai School of Medicine

This talk will start with some elementary considerations from linear algebra, in particular singular value decomposition. This procedure will be viewed from a variety of different perspectives, each leading to different insights. The talk will next focus on a number of applications of the formalism to image analysis in general, and to biomedical analysis in particular. Empirical Eigenfunctions will then be used as a stepping stone for extracting information from very weak signals, and in another line of investigation it will be shown that it leads to the extraction of basic patterns of activity for problems involving dynamics. Mention will also be made of 'Low Dimensional Approximations'.

The talk will conclude with an entertaining example (I hope) of the methodology being applied to an analysis of voting patterns of the U.S.Supreme Court.

3.30 p.m., Smith Laboratory, Room 1005

Refreshments served in Smith 1094 at 3:00 p.m.