An Improved Method for Similar Handwritten Chinese Character Recognition

Abstract

Similar characters existed in Chinese character affect much on enhancing the handwritten Chinese character recognition rate. An improved method is presented in this paper, which combines structural and statistical features for similar handwritten Chinese character recognition,. Four-corner code feature that based on stroke structure is used to get the similar character set. According to the different strokes on the four corners, four-corner code feature can dispatch some characters which are similar in shape into different similar sets. Statistical hierarchy contour features are extracted from the characters in the same similar set, then support vector machine (SVM) is adopted as classifier to recognize the similar characters. This method reduces the complexity of recognizing similar characters, and the experiment results on common used 500 Chinese characters show the effectiveness.