摘要
在手写汉字中,毛刺是常见的噪声之一,笔划间的毛刺还经常出现粘连现象,这些粘连严重影响了基于轮廓和骨架的特征提取,从而影响手写汉字的识别率。该文通过定义边缘端点、边缘歧点和边缘直角点,提出了一种迭代算法消除毛刺以及毛刺粘连。实验表明,该算法是稳定的。
Glitch is a common kind of noise in handwriting Chinese characters. Glitches between strokes always join together, which can seriously affect those features extracted from contour and skeleton,thus affecting the recognition rate. Based on the definitions of edge end point, edge fork point and rectangular edge point,an iterative algorithm is proposed to eliminate glitches and glitch joins in handwriting Chinese characters in this paper. Experiments show that the algorithm is stable and effective.
出处
《南京理工大学学报》
EI
CAS
CSCD
1999年第2期141-144,共4页
Journal of Nanjing University of Science and Technology
关键词
毛刺噪声
手写体汉字识别
消除法
毛刺粘连
sharp processing of image,word problems,edge,recognition,elimination
handwriting Chinese character