A support system for motion training using motion capture and acceleration sensors

This paper presents a support system for motion training for dances, sports, gestures, etc. In our previous study, we developed a prototype system for supporting motion training using motion caputure, i.e., marker tracking using a DV camera. The prototype system scores a user's motion by comparing tracking data of the user and a model with DP matching. The prototype system has a limit that it cannot accurately obtain acceleration of each part of the body because marker tracking using a DV camera cannot obtain three dimensional motion. By this limit, two different motions sometimes cannot be distinguished. For more effective motion training, in this study, we propose a system combining motion capture and acceleration sensors. We have examined the effectiveness of using acceleration sensors in motion training by comparing the proposed system with the prototype system. Experimental results have shown that two different motions can be more clearly distinguished by using the acceleration sensors.