Tracking persons in monocular image sequences

Quantitative geometric descriptions of the movements of persons are obtaine d by fitting the projection of a three-dimensional person model to consecut ive frames of an image sequence. The kinematic of the person model is given by a homogeneous transformation tree and its body parts are modeled by rig ht-elliptical cones. The values of a varying number of degrees of freedom ( DOFs; body joints, position, and orientation of the person relative to the camera) can be determined according to the application and the kind of imag e sequence. The determination of the DOFs is understood as an estimation pr oblem which is solved by an iterated extended Kalman filter (IEKF). For thi s purpose, the person model is augmented by a simple motion model of consta nt velocity for all DOFs which is used in the prediction step of the IEKF. In the update step, both region and edge information are used. Various expe riments demonstrate the efficiency of our approach. (C) 1999 Academic Press .