首页 | 本学科首页   官方微博 | 高级检索  
     

联机手写笔画特征抽取的逼近-合并算法
引用本文:郑胜林,潘保昌,赵学军,陈箫枫.联机手写笔画特征抽取的逼近-合并算法[J].计算机工程与设计,2006,27(7):1248-1250.
作者姓名:郑胜林  潘保昌  赵学军  陈箫枫
作者单位:广东工业大学,信息工程学院数字图像技术研究所,广东,广州,510643
摘    要:为了对联机手写字符识别的笔画进行精确描述,提出了一种基于字符笔画特征抽取的"逼近-合并"算法.该算法分析了字符笔画的多边形逼近,求出偏离度最小的多边形逼近,并对该多边形的边进行合并,抽取出笔画方向码,实现了联机手写字符笔画的更有效合并.该方法应用在联机手写体字符识别实验系统中,其识别率为99.13%.

关 键 词:文字识别  特征抽取  多边形逼近  模式识别  图像处理
文章编号:1000-7024(2006)07-1248-03
收稿时间:2005-01-18
修稿时间:2005-01-18

Approximation-integrate method of extracting stroke features on line
ZHENG Sheng-lin,PAN Bao-chang,ZHAO Xue-jun,CHEN Xiao-feng.Approximation-integrate method of extracting stroke features on line[J].Computer Engineering and Design,2006,27(7):1248-1250.
Authors:ZHENG Sheng-lin  PAN Bao-chang  ZHAO Xue-jun  CHEN Xiao-feng
Affiliation:Institute of Digital Image Technology, College of Information Engineering, Guangdong University of Technology, Guangzhou 510643, China
Abstract:An approximation-integrate method of extracting stroke features is presented for exactly describing strokes of online handwritten characters. The polygon approximation for character strokes is analyzed, the minimum deviation polygon approximation is founded. The strokes are integrated and the strokes codes are extracted. The integration of online handwritten character strokes is realized more efficiently. This method has been applied in the experimental system of handwritten character recognition. The test result shows that the recognition rate is 99.13 %.
Keywords:OCR
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号