云环境下基于函数编码的移动应用克隆检测

Function encoding based approach for App clone detection in cloud environment

提出了一种云环境下基于汇编函数层编码的App克隆检测方法,实现了Android应用克隆检测原型系统Pentagon。首先,针对App执行文件代码设计了二进制函数基本代码块特征提取方法;其次,提出了一种对二进制函数进行单调编码的算法,基于图形嵌入原理,融合执行流程图与函数代码基本块特征,形成每个函数代码的投影特征;最后,提出了一种高效的第三方库函数过滤方法,进一步提升了克隆检测效果。实验证明,所提方法具有97.6%的准确率,在实验条件下判定一个App克隆对只需要79 ms,能够支撑云环境下应用市场级别的大规模克隆检测。

An efficient function-based encoding scheme in the cloud environment for detecting the cloned Apps was designed,called Pentagon.Firstly,a basic block feature extraction method was proposed.Secondly,a monotonic encoding algorithm for the App function was designed,which encoded the function based on the control flow graph structure and basic block attributes.Finally,a three-party libraries filtering method was proposed by using an efficient clustering algorithm based on the function feature.Experiments verified the effectiveness of the proposed scheme.The average search time is close to 79 ms,and the clone detection accuracy achieves 97.6%.