基于N联字的汉字识别后处理研究

A Study Postprocessing Approach to Chinese Character Recognition Based on N-United-Word

为了提高汉字识别率，本文提出了在单个汉字的初级识别后，利用Ｎ联字的上下文关系，对初级识别中拒识或不确定的汉字语段作进一步确认的一种方法，阐明了Ｎ联字后处理方法的基本思想，给出了实现此方法的数据库的结构设计方案和理论算法，分析了理论上可提高的识别率，最后给出了一个Ｎ联字汉字识别后处理系统模型。

Abstract In order to increase the recognition rate or chinese characters, a postprocessing methodto the recognition or chinese characters based on contextual relation of N-united-wordhas been proposed in this paper. That is the method with which we can further determinedthe chinese characters that is not determined or refuese to recognize in the priliminary recognized period. First, the main idea and theoretical foundation of this method has been expounded. Second, the structure or the database of the postprocessing based on2-united-word has been discussed and the theoretical algorithm of the postproccssing ofchinese chatacters recognition has been given out.Then, the recognized rate increased intheory has been analyzed. Finally, a system model of N-united-word postprocessing hasbeen given out.