The parameter optimization of the pulse coupled neural network for the pattern recognition

Abstract

The pattern recognition using Pulse Coupled Neural Network (PCNN) had been proposed. In conventional studies, the parameters in the PCNN are used to be defined empirically and the optimization of parameters has been known as a remaining problem of PCNN. In this study, we show a method to apply the real coded genetic algorithm to the parameter optimization of the PCNN and we also show performances of pattern recognition by the PCNN with learned parameters. © 2010 Springer-Verlag Berlin Heidelberg.