Theory of Invariant Algebra and its Applications in Pattern Classification

Dr. Firooz Sadjadi

Lockheed Martin Corporation

Date: Wednesday September 7, 2005

Time: 12:20 p.m. - 1:10 p.m.

Location: 322 Votey

Abstract

Machine classification of objects and patterns independent of size, orientation, position in the field of view, color and emissivity is a difficult and important problem in robotics vision, image analysis, and automatic target recognition fields. In this talk we explore the theory of invariant algebra to develop techniques that can be used to provide solutions for this problem. Studying the intrinsic properties of polynomials, that remain undisturbed under changes of variables, forms the domain of this theory. Algebraic invariants of binary and ternary quantics are used to obtain features that remain unchanged when the object undergoes linear geometrical and spectral/material transformations. Empirical examples of the use of this approach on real and synthetic will be provided.

Speaker bio: Dr. Sadjadi Received the BSEE from Purdue University in 1972, MSEE in 1974, and the DEE in 1976 from the University of Southern California (USC). He completed his postgraduate studies for a PhD at the University of Tennessee, Knoxville. He worked at the USC Image Processing Institute and the Image and Pattern Analysis Laboratory of the University of Tennessee, and was a consultant to the Oak Ridge National Laboratory. During 1983-1993, he was with the Honeywell systems and Research Center (SRC) as a principal research scientist. In 1993 he joined the Lockheed Martin Corporation as a research staff scientist. His interests are in theoretical and experimental research related to signal and image processing, pattern recognition, target tracking and information fusion. He was the Guest Editor for the Optical Engineering Journal Special Issues on “Performance Evaluations of Signal and Image Processing Systems,” February 1991, “Automatic Target Recognition,” December 1992, and for the IEEE Computer Special Section on “Applications of Computer Vision,” July 1997.  He currently serves on the Editorial Board of the Information Fusion Journal. He is the author of more than 150 publications, and holds 8 US and International Patents. He is the editor of several books, Automatic Target Recognition Systems (2000), Sensor and Data Fusion (1996), Performance Evaluations of Signal and Image Processing Systems (1993), and a forthcoming book The Physics of Automatic Target Recognition to be published by Springer. He has been active in organizing and leading a number of scientific meeting and organizations; He has chaired the annual SPIE Automatic Target Recognition Conference for the past 15 years and currently serves as a Chairman of Automatic Target Recognition Technical Working Group and the Twin Cities chapter of the IEEE Aerospace and Electronic Systems Society. He has served on a number of governmental and academic research panels including the National Science Foundation.  He has received a number of technical awards including the Honeywell Technical Achievement Award, and the Lockheed Martin Mission Success Awards. He is a Fellow of the International Society for Optical Engineering (SPIE), is a Senior Member of the IEEE, and is a member of Sigma Xi, and Optical Society of America.