Appearance similarity evaluation for Android applications

Android, one of the most popular mobile operating system in recent year, attracts device users by its numerous applications. Google built Google Play Store as its official Android market for Android users to search and install apps. However, with the increasingly growth of the number of apps in Android marketplaces, some undesirable apps began to turn up. Plagiarisms and malware are two main kinds of those undesirable applications. For plagiarisms, they pirate other developers´ work. As for malware, one of the most common way of spread is inducing users to install by their graphical user interface (GUI) similarity. Therefore, in this paper, we put forward a method to detect plagiarisms or malware which conceal themselves as a legitimate pre-existed app in Android marketplaces by their appearance. The main method we design is to compare the GUI similarity among Android apps and pick up some apps with high similarity on their appearance. In detail, we extract some features of apps and compute their similarity by their feature vectors. At last, we evaluate our design with 2,000 apps in both official and alternative Android marketplaces to find out such appearance similar apps in our dataset.