作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (2): 200-202. doi: 10.3969/j.issn.1000-3428.2011.02.069

• 人工智能及识别技术 • 上一篇    下一篇

基于形状上下文的复杂验证码识别算法

贺 强,晏 立   

  1. (江苏大学计算机科学与通信工程学院,江苏 镇江 212013)
  • 出版日期:2011-01-20 发布日期:2011-01-25
  • 作者简介:贺 强(1984-),男,硕士研究生,主研方向:模式识别,图像处理;晏 立,教授
  • 基金资助:
    国家科技型中小企业技术创新基金资助项目(09C26213 203797)

Recognition Algorithm of Complicated CAPTCHA Based on Shape Context

HE Qiang, YAN Li   

  1. (School of Computer Science and Communication Engineering, Jiangsu University, Zhenjiang 212013, China)
  • Online:2011-01-20 Published:2011-01-25

摘要: 使用改进的形状上下文方法对复杂验证码进行识别,采用整体识别方法,不对图片进行切割,使用半极坐标圆进行建模的方式,解决2个字符连接处像素点建模互相干扰的问题。设计并实现复杂验证码识别算法,并与简单验证码进行比较。实验结果证明,复杂验证码识别算法能对字符粘连的复杂验证码进行识别。

关键词: 验证码, 形状上下文, 整体识别

Abstract: This paper uses the improved shape context to recognize the complicated CAPTCHA. Through the holistic recognition method, without cutting the picture into pieces, but adopting semi-polar coordinates to model, it solves the problem that pixel points between two adhered characteristics interfered each other when modeling, designs and realizes its arithmetic method, and makes a contrast with other simple CAPTCHA’s identifying process. Experimental result shows the method put forward in this paper can recognize complicated CAPTCHA.

Key words: CAPTCHA, shape context, holistic recognition

中图分类号: