网络与信息安全学报 ›› 2017, Vol. 3 ›› Issue (2): 31-38.doi: 10.11959/j.issn.2096-109x.2017.00122

• 学术论文 • 上一篇    下一篇

基于页面敏感特征的金融类钓鱼网页检测方法

胡向东1(),刘可1,张峰2,林家富1,付俊2,郭智慧2   

  1. 1 重庆邮电大学自动化学院,重庆 400065
    2 中国移动通信有限公司研究院,北京 100033
  • 修回日期:2016-12-24 出版日期:2017-02-01 发布日期:2017-02-10
  • 作者简介:胡向东(1971-),男,四川广安人,博士,重庆邮电大学教授,主要研究方向为网络化测控及其信息安全、物联网与智慧空间安全、复杂系统建模仿真与优化。|刘可(1992-),男,重庆人,重庆邮电大学硕士生,主要研究方向为物联网安全。|张峰(1977- ),男,湖北孝感人,博士,中国移动通信有限公司研究院高级工程师,主要研究方向为网络与信息安全技术应用。|林家富(1989-),男,四川成都人,重庆邮电大学硕士生,主要研究方向为物联网安全。|付俊(1979-),男,湖北松滋人,中国移动通信有限公司研究院研究员,主要研究方向为网络与信息安全方案设计、安全标准制定以及黑客攻防对抗技术。|郭智慧(1986-),男,河北张家口人,中国移动通信有限公司研究院研究员,主要研究方向为网络欺诈治理。
  • 基金资助:
    教育部—中国移动联合研究基金资助项目(MCM20150202);重庆市教委科研基金资助项目(KJ1602201)

Financial phishing detection method based on sensitive characteristics of webpage

Xiang-dong HU1(),Ke LIU1,Feng ZHANG2,Jia-fu LIN1,Jun FU2,Zhi-hui GUO2   

  1. 1 School of Automation, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
    2 Research Institute of China Mobile, Beijing 100033, China
  • Revised:2016-12-24 Online:2017-02-01 Published:2017-02-10
  • Supported by:
    The Joint Research Foundation of the Ministry of Education of the People's Republic of China and China Mo-bile(MCM20150202);The Science and Technology Project Affiliated to Chongqing Education Commission(KJ1602201)

摘要:

提出一种基于页面敏感特征的金融类钓鱼网页检测方法,通过获取网页超文本标记语言特定标签中的文本信息,利用适合中文的多模式匹配算法(AC_SC, AC suitable for Chinese)匹配出敏感文本条数,计算出敏感文本特征值;定位截取网页的logo图像,采用PCA-SIFT算法提取图像特征,并与预先建立的网页logo图像库进行匹配,计算出logo图像相似度;基于文本特征值和图像相似度实现对金融类钓鱼网页的判定。实验结果表明,该方法具有很强的针对性和时效性,并能取得不低于97%的召回率。

关键词: 金融网页, 敏感特征, 文本特征值, 图像相似度, 钓鱼检测

Abstract:

A financial phishing detection method based on sensitive characteristics of webpage was proposed, which acquired sensitive text information of specific hypertext markup language tags and computes sensitive text eigen-value. The method matches number of sensitive text using multiple pattern matching algorithm AC_SC (AC suitable for Chinese). Then, the method locates and cuts logo image of webpage, and utilizes PCA-SIFT algorithm to extract image features and match features with library of webpage logo which was established beforehand. Meanwhile, it calculates similarity of two logo image. Finally, the decision can be concluded by the text eigenvalue and image similarity. It shows that the method is better in pertinence and timeliness according to experiment, and achieves no less than 97% detection accuracy.

Key words: financial webpage, sensitive characteristic, text eigenvalue, image similarity, phishing detection

中图分类号: 

No Suggested Reading articles found!