Object tracking algorithm based on HSV color histogram and block-sparse representation

Sparse representation has been applied in tracking rapidly. However, most of these methods use the holistic model and grayscale image which result in insensitiveness to color information and ineffectiveness to deformation of non-rigid object. In this paper, we propose a robust and effective tracking method (SRH) based on the block-sparse representation (local information) and HSV color histogram (spatial information). This method not only keeps the merits of block-sparse representation in handling illumination changes and occlusions, but also adds the object color resolution and is not subjected to similar colors interference. It improves the accuracy and efficiency of tracking result. In addition, we calculate the weight of block-sparse representation and HSV color histogram and introduce a fusion method to fuse these two data. Furthermore, the occlusion handling reduces the influence of heavy block, the update strategy guarantees the tracker to adapt to the complex background and morphological changes of target which enhances the reliability of this method. Experimental results compared with several state-of-the-art trackers on challenging sequences demonstrate the robustness and accuracy of the proposed tracking algorithm.