Vision based Interface for Computer Games

摘要

In this paper, we propose a new vision based interface that allows users to play computer games with intuitive gestures. Proposed system detects the user's gestures from continuous image sequence and translates them into the corresponding game commands. This system consists of three main modules: body parts localization module, pose classification module and gesture recognition module. Firstly, body part localization module determines the locations of body parts such as face and hands automatically using the skin-color model. Subsequently, pose classification module classifies the positions of detected body parts in a frame into a pose according to Euclidean distance between the input positions and predefined pose. Finally, gesture recognition module extracts a sequence of poses corresponding to the gestures from the successive frames, and translates that sequence into the game commands using a HMM. In this HMM, some states are closely related with certain gestures, and when the HMM reaches that states a gesture related with that state is detected and recognized simultaneously. Each gesture has at least one such state and we call them as "distinctive states". To assess the effectiveness of the proposed interface, it has been tested with popular computer game, Quake II, and the results demonstrate reasonable overall gesture detection accuracy.