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 ›› 2009, Vol. 35 ›› Issue (15): 209-211. doi: 10.3969/j.issn.1000-3428.2009.15.073

• 人工智能及识别技术 • 上一篇    下一篇

基于颜色与角点特征的图像垃圾邮件识别算法

万明成,耿 技,程红蓉,王 勇   

  1. (电子科技大学计算机科学与工程学院,成都 610054)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2009-08-05 发布日期:2009-08-05

Image Spam Identifying Algorithm Based on Color and Corner Feature

WAN Ming-cheng, GENG Ji, CHENG Hong-rong, WANG Yong

  

  1. (School of Computer Science and Engineering, University of Electronic Science and Technology, Chengdu 610054)
  • Received:1900-01-01 Revised:1900-01-01 Online:2009-08-05 Published:2009-08-05

摘要: 垃圾邮件制造者将垃圾信息嵌入图像中,使基于文本内容的反垃圾邮件系统失效。对垃圾邮件图像的特点深入分析后,提出一种垃圾邮件图像识别算法。垃圾邮件图像多为计算机合成图像,其颜色不如自然图像丰富,且因含有大量文字导致图像中角点角度值分布呈现出一定的规律性。针对此问题选用颜色和角点特征并结合支持向量机分类算法来识别垃圾邮件图像。实验结果表明,该算法对真实垃圾邮件图像的识别精确率超过98%。

关键词: 图像垃圾邮件, 颜色特征, 角点特征, 支持向量机

Abstract: Spammers embed spam message into images and failed many text-based anti-spam systems. This paper proposes an effective method to discriminate the spam images by analyzing the features of image. Most of spam images are generated by computer, which are not as rich in colors as nature photos. Besides, as contained many texts, the images had certain regularity in corner angle distribution. The algorithm extracts some features of color and corner of image, and identifies spam image by a SVM classifier. Experimental result on a real word data shows that the accuracy rate of the proposed algorithm is more than 98%.

Key words: image spam, color feature, corner feature, Support Vector Machine(SVM)

中图分类号: