ABSTRACT
ABSTRACT: Object tracking is one of the most important research directions in the image processing. However, most existing tracking algorithms only consider two adjacent frames at a time and ignore the continuity of the whole sequence. When the object is occluded, it is difficult to locate the object in the right trajectory after the occlusion. In this paper, an object tracking algorithm in the low-rank representation has been proposed to cope with the occlusion. We decompose the observation sequence into a low-rank matrix, which is composed of the moving object, and a sparse support matrix, which is composed of the occlusion. We use the variation between two adjacent frames to locate the moving object in the sequence, and the sparse support matrix to cope with the occlusion. Several experiments on real video sequences indicate that the object tracking algorithm in the low-rank representation works effectively under conditions of the occlusion.
