Detect Android Malware Variants Using Component Based Topology Graph

摘要: Smartphone has experienced explosive growth recently. At present, Android system is the most popular mobile platform and attracts lots of developers as well malware authors. In order to evade detection, authors often apply obfuscation techniques morph malware. Since traditional detectors are based on pure syntax, they may fail detect obfuscated variants. We present a novel signature, topology graph components, which could model malicious payloads properly resist against common used by hackers. performe stress test security tools provided Virus total with ten kinds families from Malware Genome Project. Unfortunately, result not optimistic that samples tools. Nevertheless, 86.36% we tested caught our detector tolerable false positive. The evaluation demonstrates approach able variants generated techniques.