The research on features of large deformation offline handwritten Chinese character recognition

The offline handwriting recognition is one of the difficulties of pattern recognition, feature extraction results directly affects the recognition rate. Firstly, various existing structural features and statistical features will be analyzed, and feature fusion and feature extraction methods will introduction briefly. Finally, a method for feature extraction is presented, which is developed based on the multi-resolution theory and feature fusion theory. It will use these theories generate different feature and use fusion methods to generate new character feature, thereby improving the recognition rate.