Feature Extraction and Identification of Handwritten Characters

Abstract

Handwritten character recognition techniques are quickly developed during past years. In this paper, we summarize both feature extraction and classification recognition methods of off-line handwritten characters recognition. The feature extraction methods mainly include two classes: statistic features and structure features, and the classification methods usually covers pattern matching, statistical classifier, neural network classifier, etc. These techniques are analyzed and comprised, and the technology trends of handwritten character recognition are also discussed.