A robust Cam Shift tracking algorithm based on multi-cues fusion

A novel robust tracking algorithm based on multi-cues fusion under Cam Shift framework is proposed in this paper. The color and edge cues are applied to describe the object and an adjustable feature update model is incorporated into Cam Shift tracking process. The weights of the color and edges are tuned adaptively according to the probabilistic model in the tracking process, which enhances the tracking robustness. The experiment results show that the proposed algorithm could track the object robustly and accurately.