3D 모바일 아바타를 위한 적응적 계층 동작 저작

초록

Design of avatars which represents human motion is widely used in movie, computer game and virtual reality industry. In order to get the realistic human posture, motion data collection and database construction are required by using sensors such as motion capture device. Natural motion of human avatar can be achieved by using the massive amount of human posture data. However it is not easy to implement the real time interactive mobile avatar, since it requires complex calculation of those acquired posture data. This paper suggests a flexible, adaptive and hierarchical mapping of various avatar motion by using intuitive classification of posture data set through the pattern matching. Motion input sensor including gyro is used to track the user input and the SVM(support vector machine) is applied to classify the input motion.

키워드