Model Retargeting Motion Capture System Based on Kinect Gesture Calibration

The exorbitant cost required to implement a Motion Capture (Mocap) System is of tremendous concern for most Mocap producers. This research is to develop a novel approach for human motion capture that is independent of advanced markers (optical or inertial) and is cost-effective for small-scale production. Utilizing a Rig retargeting motion approach, this study implemented a system that requires a single depth sensor and free animation software to track the joint coordinates of human performance to record motion in real-time. The high precision of human tracked data mapped to Rigged Armatures proved significant in the implementation of our system. The experiment results indicate that high precision motion transfer from human actors to 3D models is conducted. The cost of equipment required for implementation is estimated to be 90% lower than the required cost of a premium motion capture setup.