This paper presents new recognition mehtod for handwritten Chinese characters. A number of integrated morphological iamge information based on complex transformed iamges and subareas is used to get good recognition rate. Similarity by morphological features are used for recognition. Complex iamges are obtained by the distance transform and the nonlinear normalization. Instead of selective use of complex image information and integrated iamge information, general similarity of each similarity based on complex images, subareas and integrated image information are used to get good recognition. General similarity is based on deviation value of each similarity. We obtained the best recognition rate 99.71
展开▼
