手写体汉字识别系统

RECOGNITION SYSTEM OF HANDWRITTEN CHINESE CHARACTERS

本文对手写体汉字提出首尾笔划特征表示法,并以此为编码信息;加上以各字形的二维DFT为该字形的总体外形特征,由此构成手写体汉字数据库. 在识别过程中,针对手写体的模糊特点,引入程度函数和不精确推理结构及其相应的算法.

In this paper, how to indicate the characteristic of starting and ending strokes of a Chinese character is presented. Using these characteristics, coding information, can be obtained. Two-dimensional Fourier transform of the fonts are used as the features of the characters. According to these a database for handwritten Chinese characters can be constructed. A architecture for inaccurate inference is constructed and the relevant algorithm is given.