3-D human motion sensing by artificial retina chips

We aim at capturing 3-D human motion from stereo images in real time without any attached markers. Our strategy comprises two tasks: identification of body parts and motion assignment to them We used a pair of artificial retina chips for implementing stereo vision. In this paper, we introduced three algorithms. The first algorithm is the 3-D motion extraction based on optical flow analysis, and we showed its performance with 3-D CG. The second algorithm is to localize an image region including a human head by frame subtraction. The final algorithm is to extract the head rotation in depth. We successfully dicriminated four kinds of head rotations. To show the performance, we built a demo system and realized that each rotation could be reflected on a head CG model.