Theories of Pattern Recognition

Abstract

This chapter discusses the theories of pattern recognition. Automatic pattern recognition may best be defined as machine identification of patterns. The types of patterns are of visual, aural, and electromagnetic origin. The recognition of patterns can only be affected on the basis of differences in the characteristics of the patterns. In general, these characteristics cannot be clearly delineated by the machine designer and therefore, the heart of most realistic pattern recognition problems is the use of learning samples in formulating the decision procedure employed by the machine. For this reason, the study of automatic pattern recognition involves the study of machine learning, and machines that perform pattern recognition must invariably be adaptive. In an attempt to build an electronic system capable of performing pattern recognition and learning, engineers are naturally intrigued by the ease with which these functions are accomplished in even the simplest biological systems.