Analysis of Human Motion in Image Sequence using 3D Model and Extended Kalman Filter

W.W. Lok and K.L. Chan (PRC)


human motion analysis, tracking, model-based, Kalman filter


Tracking human motion in an image sequence is a challenging problem in computer vision. It has found a wide range of applications such as visual surveillance, virtual reality, sports science, etc. This project aims to develop a model-based motion analysis system that can track human movement in image sequence with minimum constraint. No markers or sensors are attached to the subject. There is no need for the subject to wear tight clothing and occlusion will not seriously affect the tracking process. The 3D human model consists of 39 degrees of freedom (DOFs). Body parts are represented as right-elliptical cones. The projection of human model into the image is explicitly modelled. The pose of the subject in each image frame is predicted by the Kalman filter. The update step is achieved by matching the gradient and textural region information. Experiment has been carried out in tracking the human walking in the image sequence.

Important Links:

Go Back