Title :
Visual Tracking With Group Motion Approach
Author :
Arslan, Ali Erkin ; Demirekler, Mübeccel
Author_Institution :
Middle East Technical University Ankara, Turkey
Abstract :
An algorithm for tracking a single visual target is presented where visual features of target are tracked in a group tracking method. A single visual target is regarded as a group of multiple targets. The model developed allows some variations on the relative positions of the features which corresponds to some sort of non-rigidity. The approach is a combination of PDA and JPDA algorithms and using the advantages of group motion model this approach enables accurate tracking of single targets.
Keywords :
Kalman filtering; PDA; feature tracking; group tracking; multi-target tracking; visual tracking; Bayesian methods; Computer vision; Filtering; Kalman filters; Measurement standards; Personal digital assistants; Position measurement; Target tracking; Vehicle dynamics; Vehicles;
Conference_Titel :
Computer Vision and Pattern Recognition Workshop, 2004. CVPRW '04. Conference on
DOI :
10.1109/CVPR.2004.203