Motion Capture Based Motion Retargeting

Computer 3-dimensional animation has been the focus in the research area of computer graphics, and motion capture technology adds new energy to it. But motion capture technique has some flaws, such as using the device in the restricted field, noise in the capture data, single style of motion, etc. As the widely use of motion capture system, motion editing and motion retargeting technique become important. Through the motion editing and motion retargeting, we can let the motion data which we captured to fit a new environment.In this paper, we analyzed the motion capture data which is expressed by ASF file and AMC file. So we defined a 3-dimensional skeleton model by H-Anim standard, which include 30 bone segments, 28 joints and 56 degrees of freedom. Moreover, we studied the motion retargeting based on the model. In our motion retargeting algorithm, a new adaptation principle for end-effector constraint was introduced to solve foot hang-in-the-air and the footstep slip in the editing process. First, the position of the end-effector was computed according to the adaptation principle, which was to be the constraint in each frame. Then, the root position was adapted according to the constraint in every frame. Last, the unconstraint end-effector's position was adjusted. It keeps the primary joint's angles, so the new motion can keep the dynamics of the original motion. Per-frame editing and kinematics technique was used in our method and could compute in real time. Our method needed less manual interference in editing process than traditional ones.Motion retargeting platform was built using MFC and OpenGL. Now it can carry out motion retargeting and make a good preparation for the further research of motion retargeting.