Target-adjusted Kernel model for Mean-shift tracker

Kernel-based tracker shows robust performances in various object tracking technologies. Due to its robustness and accuracy, kernel-based tracker using mean-shift algorithm is regarded as one of the best ways to apply in object tracking technology in computer vision fields. However, it fails tracking when faced with a speedy object moving beyond its window size within one image frame interval time. These tracking failures are reduced with the use of target-adjusted kernel models proposed in this paper. Target-adjusted models are designed to reflect the target information that is collected during a sampling period. Same to the conventional kernel model in kernel-based object tracker, this model also contains the target information of both the color and the distance. Experimental results show that the time required for the calculations in tracking process is lessened so that causing less failure by applying this model. The target-adjusted model is designed to focus on the colors in the target object by using look-up tables. These look-up tables enable a system to reduce the size of the color bins in a model so that unnecessary trivial computations can be discarded.