电子学报 ›› 2019, Vol. 47 ›› Issue (9): 1913-1918.DOI: 10.3969/j.issn.0372-2112.2019.09.014

• 学术论文 • 上一篇    下一篇

基于资源签名的Android应用相似性快速检测方法

张鹏1,2, 牛少彰1, 黄如强1   

  1. 1. 北京邮电大学智能通信软件与多媒体北京市重点实验室, 北京 100876;
    2. 宁夏大学信息工程学院, 宁夏银川 750021
  • 收稿日期:2018-05-28 修回日期:2018-08-04 出版日期:2019-09-25
    • 作者简介:
    • 张鹏 男,1980年生,博士生,副教授,研究方向为移动互联网、软件保护等.E-mail:longbow27@163.com;牛少彰 男,1963年生,博士,教授,博士生导师,研究方向为信息隐藏、移动互联网等.E-mail:szniu@bupt.edu.cn;黄如强 男,1995年生,硕士生,研究方向为智能终端安全等.
    • 基金资助:
    • 国家自然科学基金资助项目 (No.U1536121,No.61370195)

A Fast and Resource-Based Detection Approach of Similar Android Application

ZHANG Peng1,2, NIU Shao-zhang1, HUANG Ru-qiang1   

  1. 1. Beijing Key Lab of Intelligent Telecommunication Software and Multimedia, Beijing University of Posts and Telecommunications, Beijing 100876, China;
    2. College of Information Engineering, Ningxia University, Yinchuan, Ningxia 750021, China
  • Received:2018-05-28 Revised:2018-08-04 Online:2019-09-25 Published:2019-09-25
    • Supported by:
    • National Natural Science Foundation of China (No.U1536121, No.61370195)

摘要: 由于盗版Android应用(Android Application,简称APP)通常保持着与正版APP相似的用户体验,因此本文提出一种基于资源签名的APP相似性快速检测方法.该方法将APP的资源签名视为字符串集合,利用计算任意一对APP资源签名集合的Jaccard系数判断两者的相似性.为了避免遍历全部的APP对,该方法将MinHash和LSH(Locality Sensitive Hashing)算法的思路引入其中,通过从APP集合中挑选候选对并对候选对进行检验的方式获得最终的检测结果.由于挑选候选对的方式将大量相似性较低的APP对排除在外,因此该方法可以明显地提高APP相似性的检测速度.实验结果表明,该方法的检测速度比现有方法FSquaDRA提高了大约30倍,而检测结果与FSquaDRA几乎完全相同.

关键词: APP相似性, 资源签名, MinHash, LSH, Jaccard系数

Abstract: Since pirated Android applications (APPs for short) usually maintain a similar user experience to original APPs, a fast APP similarity detection approach based on resource signature has been proposed. In order to determine the similarity of a pair of APP, the approach calculates the Jaccard coefficient of resource signature sets of them because a set of resource signatures can be treated as a set of strings. With the help of the MinHash and LSH (Locality Sensitive Hashing) algorithm, it can avoid the traversal of all APP pairs by selecting candidate pairs from the APP set and verifying them at last.Because the procedure of selecting candidate pairs excludes a large number of APP pairs with lower similarity, this approach can significantly improve the detection speed of APP similarity. The experimental results show that the detection speed of this approach is about 30 times higher than the existing approach FSquaDRA while the detection result is almost identical.

Key words: APP similarity, resource signature, MinHash, LSH, Jaccard coefficient

中图分类号: