Motion Fusion Technique Based On Time And Space Warping

Computer animation is an important branch of computer graphics, which has become an important method of media production. Because the traditional animation method based on manual key-frame has low-efficiency, it cannot meet the growing demand. Motion capture technique captures human motion using specific instrument, which increases the efficiency and sense of reality. The motion capture data is usually fit for the original role, so it is hard to apply to new environment because of its limitation. In order to increase the reusability of motion captured data and meet new environment, role or style, motion capture data need to be edited. Motion fusion is a key technique of motion editing. This paper implements an intensive study of motion fusion techniques.This paper proposes a motion fusion method without manual intervention and suggests a method to identify the motion cycle based on joint movement. The cycle is calculated by the analysis of the motion capture data and the angle between the knee and the hip nodes. High-quality motion blending animation is generated after the steps of time and space warping, interpolation, and constraints reconstruction. In this paper, another new method based on multi-resolution model is proposed. It calculates frame distance by multi-resolution model; meanwhile it improves the original slope limit and adjustment window limit setting and proposes a method based on motion cycle that is to set the adjustment window limit. Multi-resolution model can calculate frame distance effectively. The improved slope limit and adjustment window limit would give restrictions to dynamic timewarping and generate good timewarping curve. The result shows that the proposed algorithms can make motion fusion fast and efficient which can generate high quality animation.This paper realizes a motion edit system which can render scene and show human skeleton model. User can make motion editing operation including motion fusion.