3D Motion Analysis of Human Arm Based on Multi-Constraints Fusion
基于多约束融合的人手臂三维运动分析

Motion estimation of human from a single or a monocular sequence of image is an important research field in Computer Vision and Image Processing. In this paper, based on constraint fusion, we proposed 3D motion estimation from the monocular perspective images without any special markers. We consider the arm as stick model with three joints, and the link between every two neighboring joint points is regarded as rigidity. Based on the knowledge of anatomy and kinematics, we exert coplanar constraint, rigidity constraint, smooth motion constraint on this model. By pre processing the image sequence, we semi automatically yielded 2D correspondences of joint points of arm among images. Then, by constraint fusion, 3D relative structure of arm can be determined in a scale factor. The experiment with real images is included. In order to testify the validity of the proposed method, we compare the 3D motion of human arm obtained by the motion capture system with that obtained by our method. The result shows that the method is very efficient.