Parallel combination of correlation and compressive trackers on visual object tracking

Even though numerous state-of-the-art methods have been developed, there still hardly exists single approach that can handle all the scenarios. A novel parallel combination of discriminative scale space tracker (DSST) and compressive tracker is proposed in this paper. In addition, Peak to Sidelobe Ratio judgment mechanism is addressed to further improve the robustness of DSST in long-term occlusion. We compare our method with 13 state-of-the-art tracking algorithms. The evaluation results show that our approach has significant improvement over state-of-the-art approaches.