作者：门光福;

机构：宁夏大学数学计算机学院;

中文关键词：西夏字;;OCR;;多级分类

英文关键词：xixia characters;OCR;multi-stages classification

中文摘要：字符识别是模式识别领域的一个传统课题,汉字和古文字的识别是中文信息处理领域的一个重要研究课题.采用了一种多级分类的方法对印刷体西夏文字进行识别,在6000余西夏文字中随机选取300字做测试,平均识别速度达63字/s,识别率可达到89%以上,对西夏文献中的西夏文字的识别奠定了基础.

英文摘要：The characters recognition is a conventional project of model recognition domain.Both chinese characters and ancient characters recognition are an important project.A multi-stages classification for machine printed xixia characters is proposed,and the experiment is used to show the recognition speed and rate can reach 63 word per second,89 percent.The investigation has built a solid theory foundation for the research and development of xixia characters recognition in xixia documents.

[中文刊名]：高师理科学刊 [出版日期]：2010-07-30 [CN]：23-1418/N [ISSN]：1007-9831

下载全文 下载CAJ阅读器