Automatic color space switching for robust tracking

This paper introduces an algorithm to automatically and continuously select the most appropriate color space to use in order to improve the performances of visual tracking. Eight color spaces are tested, and the Mean-Shift (MS) tracker is considered. The selection of the colorspace is made using an evaluation criterion based on the quality of the weights involved in the MS tracking, and implicitly on the good separability between the target and its close background. Experiments on real sequences show the impact of the color space on tracking performances and the relevancy of the proposed selection criterion.