Abstract
In this paper, we extend per-frame inverse kinematics technique for retargeting motion, and propose end-effector correction method to increase the similarity of the retargeted motion.Per-frame inverse kinematics uses end-effector of original skeleton in the motion to obtain joint angles of target skeleton. Retargeting by the trajectories of source end-effector can guide limbs to reach validly position, and preserve important aspects of the original motion. However, the retargeted motion may not achieve required quality of similarity. In this study we propose end-effector correcting mechanisms. Base on these mechanisms, two retargeting algorithms, which can perform similar motion without violating the given constraints are introduced. Moreover, empirical tests are given to demonstrate the effectiveness and efficiency of our algorithms.