Title :
Robust approach to tracking human motion in real scenes
fDate :
11/24/1994 12:00:00 AM
Abstract :
A computational system for estimating the positions and the movements of human bodies interacting in real scenes is described. The approach consists of three steps: mathematical modelling of the human body; extraction and matching of corresponding 2D features from a sequence of images; estimation of the motion and structure parameters by using Kalman filtering
Keywords :
Kalman filters; feature extraction; motion estimation; parameter estimation; physiological models; tracking; 2D feature extraction; 2D features matching; Kalman filtering; computational system; human body modelling; human motion tracking; image sequences; mathematical modelling; motion estimation; real scenes; structure parameters estimation;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19941410
