Pattern recognition: From Classical to Modern Approaches

Presents recent developments in the classical and modern hybrid methodologies currently being applied in pattern recognition. The 21 chapters deal with decision theoretic classification using the statistical approach, neural networks, the fuzzy set approach, the use of genetic algorithms, and soft computing. Topics include Bayesian approaches for unsupervised classification, networks of spiking neurons in data mining, adaptive segmentation techniques for hyperspectral imagery, and writing speed and writing sequence invariant on-line handwriting recognition.