NOVEMBER 23, 1999, 1:00
Dr. Venu Govindaraju

CEDAR, State University of New York at Buffalo
Adaptive Pattern Recognition

ABSTRACT

Combination methods for classifiers has been a topic of great interest to the Pattern Recognition community. Several combination strategies of serial and parallel combination have been developed to address specific application needs. However, a general combination theory has been lacking. We will present the research issues that must be addressed in order to develop such a general theory in the context of handwritten word recognizers. One of the issues is the notion of adaptive word recognition. The idea is that a recognizer must dynamically adapt to the classes (lexicon) it is discriminating. It follows that our first need is for a character recognizer that can adapt to its operating environment. Our presentation will focus on adaptive character recognition. We will present a method where features at different resolutions, from coarse to fine-grained, are implemented by means of a recursive classification scheme. Typically, recognizers have to balance the use of features at many resolutions (which yields a high accuracy), with the burden on computational resources in terms of storage space and processing time. We present a method that adaptively determines the degree of resolution necessary in order to classify an input pattern. This leads to optimal use of computational resources. The adaptive character recognizer dynamically adapts to factors such as the quality of the input pattern, its intrinsic similarities and differences from patterns of other classes it is being compared against and the processing time available. Furthermore, the finer resolution is accorded to only certain "zones'' of the input pattern which are deemed important given the classes that are being discriminated. Experimental results support the methodology presented. ------------- Biography: Venu Govindaraju received his PhD in computer science from the State University of new York at Buffalo in 1992. He has coauthored more than 80 technical papers in various International journals and conferences and has one US patent. He is currently the associate director of CEDAR and concurrently holds the research associate professor ship in the department of Computer Science and Engineering, State University of New York at Buffalo. He is the associate editor of the Journal of Pattern Recognition and the area chair of the IEEE SMC technical committee for pattern recognition. Dr. Govindaraju has been a co-principal investigator on several federally sponsored and industry sponsored projects. He is presently leading multiple projects on postal applications. He is a senior member of the IEEE.