Tracking occluded objects using chromatic co-occurrence matrices and particle filter

In the computer vision field, many real-world applications are based on detecting and tracking moving objects. One of the most important challenges in these applications is tracking occluded objects. Actually, when two or multiple objects occlude, the used tracking system suffers from information loss which negatively influences its tracking performance. The present paper introduces a new method to overcome this problem using only one target image and without any classification or learning phase. Indeed, a tracking system is established by combining the chromatic co-occurrence matrices and the particle filter in order to evaluate the occluded target position. The qualitative and quantitative studies show that the results obtained by the proposed approach are very competitive in comparison with several state-of-the-art methods.