Computer Engineering and Applications ›› 2011, Vol. 47 ›› Issue (10): 158-160.

• 图形、图像、模式识别 • Previous Articles     Next Articles

Single character font recognition of character dependent

CAI Kaiwei,LV Xiaoqing,LI Monan,TANG Yingmin   

  1. Institute of Computer Science & Technology of Peking University,Beijing 100871,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-04-01 Published:2011-04-01

文字相关的单字符字体识别方法

蔡凯伟,吕肖庆,李沫楠,唐英敏   

  1. 北京大学 计算机科学技术研究所,北京 100871

Abstract: Chinese characters have various fonts,and their structures show significant differences.OCR technology emphasizes on character recognition while font recognition does not get enough attention.This paper proposes character dependent font recognition approach,which makes use of prior information of character.The features are character discriminative structure features,and vector space model is used to obtain similarity measurement.Experimental results on 66 common used fonts show that average precision of each font is preferable.

Key words: font recognition, vector space model, Chinese character feature

摘要: 汉字具有丰富的字体类型,并且不同的字体在汉字结构上有显著的不同,现在的OCR技术侧重字的识别,而对字体识别的关注较少。提出文字相关的单字符字体识别方法,利用文字相关的先验信息及字体结构特征,对字体的相似性度量采用向量空间模型,并针对常用66款简体字进行实验,得到了较好的平均识别率。

关键词: 字体识别, 向量空间模型, 汉字特征