Statistical pattern recognition: A review

The primary goal of pattern recognition is supervised or unsupervised class ification. Among the various frameworks in which pattern recognition has be en traditionally formulated, the statistical approach has been most intensi vely studied and used in practice. More recently, neural network techniques and methods imported from statistical learning theory have been receiving increasing attention. The design of a recognition system requires careful a ttention to the following issues: definition of pattern classes, sensing en vironment, pattern representation, feature extraction and selection, cluste r analysis, classifier design and learning. selection of training and lest samples, and performance evaluation. In spite of almost 50 years of researc h and development in this field, the general problem of recognizing complex patterns with arbitrary orientation, location, and scale remains unsolved. New and emerging applications, such as data mining. web searching, retriev al of multimedia data, face recognition, and cursive handwriting recognitio n, require robust and efficient pattern recognition techniques. The objecti ve of this review paper is to summarize and compare some of the well-known methods used in various stages of a pattern recognition system and identify research topics and applications which are at the forefront of this exciti ng and challenging field.