JISE

    This paper presents a model based on-line recogntion system by using an AI based method for recognizing square handwritten Chinese characters with a large alphabet size. A knowledge model of Chinese characters is developed as the template feature to represent the reference patterns of the system. It contains the hypothetical knowledge of handwriting variations, including stroke-order deviations and stroke-number deviations. For pattern pair matching, a matching tree is constructed by combining the knowledge of reference characters and the primitive stroke-sequence of unknown characters together. With the tree, a similarity measure function is defined to indicate similarity degree. The evaluation of the function is obtained by utilizing A* algorithm based matching. Experimental results are based upon the testing set of 54,010 handprinted sample characters square written by 10 pepole. The cumulative classification rate of choosing the 10 most similar characters is 98%. The results suggest that the approach of the recognition system is feasible and reasonable. The process of primitive stroke extraction is also presented.