Application Oriented Semantic Multi-touch Gesture Description Method

To make multi-touch gestures better fit applications and facilitate the recognition of multi-touch gestures, this paper presents an application oriented semantic multi-touch gesture description method. Multi-touch gestures are divided into atomic and combined gestures. The atomic gesture corresponds to a single trajectory and is represented by the shape feature of the trajectory; the combined gesture is consisted of multiple trajectories and modeled by atomic gestures with spatial, temporal and logical relationships. Semantic constraints of a specific application are added and represented by the accessible relationship between entity states and combined gestures. A description tool is developed and examples show that this method is independent of multi-touch platforms and system environments, and it can improve the recognition efficiency. It should be of significant utility to advance the use of multi touch technology.