机构地区: 华南理工大学电子与信息学院无线电与自动控制研究所
出 处: 《电路与系统学报》 1996年第3期63-67,共5页
摘 要: 汉字识别中的结构类特征具有抗变形的优点,合适做为手写体识别过程使用,但其中笔划特征的提取非常困难。本文用小波变换这一图象分析中有力工具对汉字图象分析处理之后,再进行笔划的提取,其过程简洁、明晰。 The structural features of Chinese characters are suitable for the handwwring recognition process because of their advantage of high tolerances to distortions and style variations. But the extraction of the stroke feature in them is very difficult. Wavelet transformation, which is a very powerful tool in image analyses, is used to analyse the Chinese character image in this paper, and then the process of the stroke extracion becomes concise and distinct.