论文部分内容阅读
本文叙述了一个基于改进的"汉字识别特征点方法"的高精度简、繁体印刷体汉字文本识别系统。引入特征点的方向属性,明显地提高了"汉字识别特征点方法"的汉字识别率。文中阐述了该系统各主要环节的原理。经过百万汉字真实印刷文本的严格测试,本系统汉字识别率达到97.84%。对质量较高的真实印刷文本,汉字识别率达到99%以上。
This article describes a high-precision simplified and traditional printed Chinese character text recognition system based on the improved “Chinese character recognition feature point method.” By introducing the directional attributes of feature points, the Chinese character recognition rate of “Chinese character recognition feature point method” is obviously improved. The paper describes the principle of the main aspects of the system. After millions of Chinese characters printed text of the strict testing, the system recognition rate of 97.84% Chinese characters. On the higher quality of the real printed text, Chinese character recognition rate of 99% or more.