Framework for vision-based sensory games using motion estimation and collision responses

In this paper, we present a new framework for sensory games using motion estimation and collision responses. In this framework, only one Web camera is used, and the additional marker is not necessary. User's actions are extracted by analyzing consecutive frame differences. Collisions between the virtual objects and the user are detected by colliding ratio and distance, and the objects respond by the estimation of the collision response in which the collision plane is constructed by analyzing collision regions. Collision of the objects with other objects or boundaries (planes) responds by using the classical methods. Using the proposed framework, we implemented three games: two educational games and one sports game. These games are processed fast enough for real-time implementation, and can be played excitingly with simple environments. In addition, various applications may be easily developed using our framework.