利用句法结构分析的联机手写体汉字识别

On-Line Recognition for Handwritten Chinese Characters using Syntax Analysis

本文提出利用句法结构分析对限制性手写体汉字进行联机识别的方法。首先,对汉字的笔划进行编码,平滑滤波和数据压缩等予处理;然后,利用其结构信息,采用模糊开关和自动机理论,建立一种新型自动机——条件有限状态自动机,以得出手写体汉字的笔划顺序码;最后,利用汉字笔划的顺序码在识别字典中检索,查找标淮汉字的机内码,从而完成整个识别过程。

Handwritten Chinese Characters recognition using syntax analysis are proposed, strokes encoding is first introduced. The smoothing operation for stroke and data contracting is then described. Using fuzzy switch concept and automat theory tomake up a new type automat-conditional finite-state automat according to thestructure features so as to recognize the handwritten Chinese characters. It is not necessary to store any standard models and to carry out complex calculation. The recognition operation can be completed fastly.