Kernel-Based Robust Tracking for Objects Undergoing Occlusion

Visual tracking has been a challenging problem in computer vision over the decades. The applications of Visual Tracking are far-reaching, ranging from surveillance and monitoring to smart rooms. Occlusion is one of the major challenges that needs to be handled in tracking. In this work, we propose a new method to track objects undergoing occlusion using both sum-of-squared differences (SSD) and color-based mean-shift (MS) trackers which complement each other by overcoming their respective disadvantages. The rapid model change in SSD tracker is overcome by the MS tracker module, while the inability of MS tracker to handle large displacements is circumvented by the SSD module. Mean-shift tracker, which gained more attention recently, is known for tracking objects in a cluttered environment. Since the MS tracker relies on the global object parameters such as color, the performance of the tracker degrades when the object undergoes partial occlusion. To avoid the adverse effect of this global model, we use the MS tracker so as to track the local object properties instead of a global one. Further a likelihood ratio weighting is used for SSD tracker to avoid drift during partial occlusion and to update the MS tracking modules. The proposed tracker outperforms the traditional MS tracker, as illustrated in the instances applied.