“Dancing icons” detection

Undoubtedly, a key feature in the popularity of smartphones is the numerous applications one can install. Frequently, we learn about applications we desire by seeing them on someone else's mobile device. A user-friendly way to obtain these particular applications could be by taking a photo of their corresponding icons as displayed on our friend's screen. We then need to develop methods for automatic detection and recognition of the icons in a screen shot. This paper suggest a method for icon detection (icon recognition is left for future work). The variety of icons (∼500K) and wallpapers makes the detection task very difficult using methods such as edge detection, contour detection or template matching. In order to bypass this difficulty, we suggest using a special feature introduced in several smart phone. When one enters the edit mode for organizing icons on the screen, the icons vibrate. Furthermore, When one moves from one set of icons' view to another, the icons slide on the screen. We use this feature to obtain a better detection of the icons on the screen.