A Novel Method of Feature Extraction and Classification for OPCCR

In optical printed Chinese character recognition (OPCCR), support vector machine (SVM) is thought to be a good classifier. However, the recognition rate of SVM depends on the features extracted and the time consumption of it is large. For this reason, we propose statistic features (SF) and local nearest neighbor SVM (LNN-SVM) to promote the recognition rate and to reduce the computational time of SVM. Experiments have been done and the results showed that SF and LNN-SVM can promote the recognition rate and reduce the computational time in OPCCR.