A feature-based neural wavelet optical character recognition system

Abstract

A hybrid character recognition system is studied using feature extraction method. The features are extracted using wavelet transform, pre-classified using a k-nearest neighbor-based neural net and subsequently post-processed using an optical correlator.