A robust vision-based technique for human arm kinematics identification
Omid Talakoub, Farrokh Janabi Sharifi · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2006
Motion, independent of forces, is described by kinematic parameters (usually using Denavit-Hartenberg convention) that have been used widely in biomechanical fields. Examples of these fields include robotics, human motion studies, and biomechanical structures' design and control; e. g. for exoskeleton and artificial human arm. A common way to precisely measure the joints movements is by using a motion capture system. Until now, the most successful motion capture technology is optical motion capture; this is due to its highly accurate measurement of small reflective markers that are attached to some relevant body landmarks. This paper addresses the problem of estimating the human arm kinematics parameters from video or captured images of human arms. We introduce a new robust framework that leads to reliable and accurate estimation of shoulder and elbow center of rotations along with the arm kinematics parameters.