48 Dynamic Object Tracking in Occlusion & Same Background in Video

In a video sequence, object tracking became difficult if the object is occluded partially and if the background is same as that of object color. Here under these two constraints object is tracked successfully by applying the partitioning of the object technique and then considering the confidences of each region. The object is partitioned such that it's surrounding is also taken into consideration. By exploiting the attention shifting among local regions of a human vision system during tracking, a robust system is designed. By matching and comparing with the highest confidences region, a weight is assigned to each partition region of an object. The algorithm is tested for the constraints of partial occlusion and same background.