Android Repackaging Detection On User Side

Mobile devices have developed rapidly in recent years,bringing astronomical amount of mobile applications but also more and more security problems,like Android repackaging.Detection of repackaged applications can prevent malicious behavious happening on user side.Many other methods aim to server side,unable to be implemented on user side.To solve this problem,this thesis proposed a repackaging detection method on user's side,which can make use of features including images and texts to do detection on the user side.The details are as follows.First,it implemented an image similarity method based on local edge feature.It used local edges of images to compute image similarity,which make it possiple to find similar applications on image level.The experimental results show that the mehod is feasible and effective.Second,it implemented a text similarity method based on multiple grained text features.It used multiple grained feature of applications'name and package name to compute the text similarity,which make it possible to find similar applications on text level.The experimental results show that the mehod is feasible and effective.Third,it implemented a repackaging detection method on user side.With analysis of repackaged applications,it proposed detection procedures and fast similar applications search,which help to do Android repackaging detection on user side fast and precisely.The experimental results show that detection based on our method is feasible and effective.