3D human kinematical motion retrieval system

Since the optical sensor device such as Motion Capture was used to develop a system which can embed skeleton into a virtual environment, analysis of motions in a 3D scene becomes a useful technology. With Motion Capture Camera and animations represented by VRML, it is possible to automatically summarize, analyze and adjust the 3D motions of a real person. In this paper, we propose a human motion retrieval system, which allows users to retrieve 3D kinematical motions. The system includes two major components for motion analysis and comparison. The first is a recognition unit of motion types which is based on Skeleton Discrimination Tree. The second unit includes motion and synchronization similarity. The comparative approach is based on mutative dynamic programming that considers the degree of the included angles of the vectors belonging to individual feature tracks. These track points include head, knee, elbow, wrist, etc. and further aggregate important feature joints of human motion. Users may provide a VRML human motion object and find the similar human motions via our system. As a result, the system can automatically retrieve similar actions. The results are tested by three professors of physical education and our master students with a good satisfaction.