Fusion study on optical correlation pattern recognition and neural networks

摘要

Abstract: From the view point of neural networks (NNs) the present problems of optical correlation pattern recognition (OCPR) are analyzed. It is shown that distortion invariant pattern recognition (DIPR) can not be performed with a single filter within the framework of OCPR. By means of fusion between OCPR and NNs some solutions to an effective optical DIPR are proposed.!4