计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (18): 188-190.

• 图形、图像、模式识别 • 上一篇    下一篇

一种SVM验证码识别算法

殷 光,陶 亮   

  1. 安徽大学 计算智能与信号处理教育部重点实验室,合肥 230039
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-06-21 发布日期:2011-06-21

Verified code recognition algorithm based on SVM

YIN Guang,TAO Liang   

  1. Key Lab of Intelligent Computing and Signal Processing of MOE,Anhui University,Hefei 230039,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-06-21 Published:2011-06-21

摘要: 设计验证码的主要目的是区分人类和计算机,用来防止网络机器人的一些恶意行为。验证码的出现也催生了一批新工种,电子商务的发展迫切需要一种推广方式来推销他们的商品,残障人士上网问题也需要迫切关注,因此许多人开始研究网络机器人技术,用来实现邮箱自动注册、群发信息、自动灌水、自动登录等功能。目前,各种类型网站系统都利用验证码阻止网络机器人入侵,从而验证码识别技术成为研究热点。基于SVM技术对图像验证码进行识别,取得了良好的效果。

关键词: 支持向量机, 序贯最小化算法, 选择工作集, 核函数, 多类分类, 验证码识别

Abstract: The usage of verified code is to distinguish between human and machine behavior,which can be used to prevent the malicious behaviors of web robot.The emergence of verified code also makes a number of new types of jobs.The development of e-commerce needs a certain way to promote its goods,and disabilities need urgent attention;thus,many people begin to study web robot technology.They hope that it can realize automatic registrations for email systems,multi-send messages,automatic flood-blogging and auto-login et al.At present,various types of web systems make use of verified code to prevent robots invasion,so the verified code recognition technology is a research hotspot.Based on SVM(Support Vector Machine),a verified code recognition algorithm is proposed,and the experimental results show that the algorithm is efficient.

Key words: Support Vector Machine(SVM), Sequential Minimization Optimization(SMO), working set selection, kernel function, multi-class classification, verified code recognition