作者: Lingguang Lei , Yuewu Wang , Jiwu Jing , Zhongwen Zhang , Xingjie Yu
DOI: 10.1007/978-3-642-37682-5_7
关键词:
摘要: Monetary theft attacks are one of the most popular attack forms towards Android system in recent years. In this paper, we present MeadDroid, a lightweight real-time detection atop Android, to hold back type attacks. An FSM monetary is constructed, based on analysis real-world Employing an FSM-based approach, with information obtained from dynamically monitoring API calls and tracking processing flow UI (User Interface) inputs, MeadDroid can detect effectively incurs only small performance overhead. addition, realized as extension Dalvik VM, transparent user, thus provide good user experience. Based prototype system, experiments conducted 195 applications. 11 applications found accuracy almost 100% through comparing results charge bill phone number used experiments. The overhead CPU-bound micro-benchmark 8.97%. Experimental demonstrate that has terms effectiveness efficiency.