An improved object tracking method based on particle filter
Nan Liang, Lei Guo, Ying Wang · 2012
The conventional particle filter uses system transition as the proposal distribution. In order to improve the performance of particle filter for target tracking, Ensemble kalman filter is proposed to construct proposal distribution for sampling particle. In the tracking process, color model and shape model are combined and updated adaptively. Experimental results show the proposed algorithm improves the stability of the object tracking and enhances the estimation accuracy compared to conventional filters.