Research On Real-time Human Animation Synthesis And Motion Retargeting Based On Kinect

Along with character animation widely used in file, game and so on, the real-time3D animation synthesis technology has become an important research filed.3D human animation synthetic technology cannot solve real-time low cost problem of human animation synthetic, no matter the traditional man-made techonoly that depending on animators’experience or3D animation techonoly that combined motion capture device, until Microsoft introduced the Kinect equipment.According to the characteristics of Kinect equipment, a series of work was carried out in the aspect of real-time human animation synthesis and motion retargeting based on Kinect. Main works are summarized as following:1) To solve the problem of search time in the process of rebuilding high-dimensional human motion data, we proposed a searchs algorithm based on the neighbor graph to find the synthesis of the two motions and an accelerative match algorithm to search K similar data.2) Extract the skeleton from the3D mesh model automatically through gathering3D connect graph using the improved algorithm. Then, match the extracted skeleton with the joint points of the human skeleton automatically using the joint mapping approach presented in this paper.3) Propose the motion retargeting based on four constraints to eliminate movement distortion such as sliding, feet dangling, embedding in the ground and so on.4) Based on above works, this paper designs a Human animation synthesis system based on Kinect and motion retargeting. It can use Kinect to synthesis real-time, high-dimension human motion data, then use motion retargeting to drive different proportions of3D mesh model.The experiments show that, the animation synthesized through this paper’s method is real-time, vivid and provided with the original motion characteristic.