一种用于汉字识别的字形编码系统设计与实现.docVIP

一种用于汉字识别的字形编码系统设计与实现 　　摘 要： 汉字字形变化多种多样，印刷体字符具有字体差异，手写字体更是没有统一的规则，识别难度非常大，当前用于汉字识别的字形编码方法大多依据字符，无法区分笔画相近的汉字。为此设计一种新的用于汉字识别的字形编码系统，介绍了汉字字形编码的原理，并给出了字形设计方案，依据编码原则，按照汉字被拆分的部件个数对汉字字形编码方案进行设计。详细介绍了汉字输入编辑器IME的结构，通过IME实现汉字的输入。依据汉字的使用频率与分布特性，通过数理统计工具设计含有汉字活动字库的操作系统，主要包括CC?DOS和MPC?DOS操作系统。实验结果表明，采用所设计系统对汉字进行识别精度较高且编码时间少、能耗低。 　　关键词： 汉字识别； 字形编码； 印刷体字符； 汉字输入编辑器 　　中图分类号： TN911?34； TP391 文献标识码： A 文章编号： 1004?373X（2017）02?0060?04 　　Abstract： As Chinese character glyph changes variously， the printed characters have the font difference， and the rules of handwriting fonts have not been unified， the identification difficulty is very big. The current font coding method for Chinese characters identification is based on characters， and unable to distinguish between similar strokes of Chinese characters， so a new glyph coding system used for Chinese character recognition is designed. The principle of Chinese character glyph coding is introduced and a glyph design scheme is given in this paper. Chinese glyph coding scheme is designed according to the principles of coding and the quantity of the dismantled parts of Chinese characters. The structure of the input method editor （IME） for Chinese characters is introduced in detail. The input of Chinese characters is achieved by IME. According to the use frequency and distribution characteristics of Chinese characters， the operating system with Chinese character activity font library was designed by means of the mathematical statistics tools， in which the CC?DOS and MPC?DOS operating systems are included. The experimental results show that the designed system′s the identification accuracy for Chinese characters is high， its encoding time is less， and its energy consumption is low. 　　Keywords： Chinese character recognition； font encoding； character in printing form； editor for Chinese character input 　　0 引 言 　　随着计算机的逐渐发展，人们希望计算机能够独立识别印刷在纸上或人写在纸上的文字[1?3]。因此，设计一种有效的字形编码系统具有重要意义，已经成为相关学者研究的重点课题，受到越来越广泛的关注[4?6]。 　　目前，有关字形编码的研究有很多，相关研究也取得了一定的成果，