摘要： 文中介绍了一种无笔序手写体汉字在线识别方法，提出了基于汉字笔画数的编码方法。整个识别过程可分为两大部分：首先进行笔画分类，然后根据笔画序列码和汉字的结构特征完成对输入手写体汉字的在线识别。由于本系统要求可以无笔序地进行汉字输入，给汉字的分类带来很大的困难。为此，在汉字结构匹配时，提出了一种结构特征搜索及排序算法，很好地解决了要求无笔序输入所带来的问题。实践证明，这种方法用于无笔序手写体汉字的在线识别是非常有效的。

关键词: 无笔序, 手写体汉字, 在线识别

Abstract: This paper describes an approach to recognition of on-line hand-written Chinese characters based on nonsequential strokes. A kind of coding method depended on strokes number is presented. The recognition is in two steps:Firstly, the strokes are classificated, then employing strokes sequence code and Chinese characters topology structure, we can finish the input hand-written Chinese characters classit1cation. Because the system requests that users can input Chinese characters based on nonsequential strokes, A large amount of difficulty was made for Chinese characters classification. We present a new algorithm for seeking structure feature when matching Chinese characters structure and solve the problem. The experiment shows that this method worked efficiently for on-line hand-written Chinese charactres nonsequential stroke recognition.

中图分类号: 

• TP391.1