A comparison of neural network models for pattern recognition

摘要

A brief survey of the existing neural network models forsignal/image processing and pattern recognition is presented. Acomparison of the back-propagation algorithm for multilayer perceptionand an adaptive sample set construction procedure offered by Nestor'srestricted Coulomb energy network is presented. A performance comparisonwith real data for ultrasonic nondestructive evaluation of materials ispresented