On sequential Monte Carlo sampling methods for Bayesian filtering
Statistics and Computing
Object Tracking with an Adaptive Color-Based Particle Filter
Proceedings of the 24th DAGM Symposium on Pattern Recognition
Segmentation based particle filtering for real-time 2d object tracking
ECCV'12 Proceedings of the 12th European conference on Computer Vision - Volume Part IV
Hi-index | 0.00 |
To improve accuracy and robustness of video target tracking, a tracking algorithm based on a new adaptive particle swarm optimization particle filter (NAPSOPF) is proposed. A novel inertia weight generating strategy is proposed to balance adaptively the global and local searching ability of the algorithm. This strategy can adjust the particle search range to adapt to different motion levels. The possible position of moving target in the first frame image is predicted by particle filter. Then the proposed NAPSO is utilized to search the smallest Bhattacharyya distance which is most similar to the target template. As a result, the algorithm can reduce the search for matching and improve real-time performance. Experimental results show that the proposed algorithm has a good tracking accuracy and real-time in case of occlusions and fast moving target in video target tracking.