联机手写体汉字识别后处理技术的研究

文中提出了一种规则和统计相结合的计算语言模型应用于联机手写体汉字识别后处理的技术，把基于统计的大词表Ｍａｒｋｏｖ语言模型与语言规则量化模型，通过词网格技术集成在一个语言解码器，这种后处理方法由３个阶段组成，词网格生成，语言解码，基于Ｃａｃｈｅ的自学习机制，语言解码器采用Ｖｉｔｅｒｂｉ搜索算法求解最优语句候选，该项技术已应用于ＨＰＣ（手持机）手写电脑的联机汉字手写体识别系统中，汉字识别率为９１．３％

A POST PROCESSING METHOD FORONLINE HANDWRITTENCHINESE CHARACTER RECOGNITION

Proposed in the paper here is a new post processing method integrating the rulebased grammar and the Markov language model for online handwritten Chinese character recognition. The Markov language model and the quantification rules model are bound to a linguistic decoder by word lattice. The post processing kernel engine consists of three stages: word lattice formation, linguistic decoder,and cache|based self|learning mechanism. The linguistic decoder adopts Viterbi search algorithm to search the best sentence hypothesis. The introduced technique has been applied to HPCs online handwritten Chinese character recognition,with a recognition accuracy rate of 91.3% achieved.