A Robust Approach Toward Face Tracking

Abstract

The mean shift algorithm has achieved considerable success in object tracking due to its simplicity and robustness. It finds local minima of a similarity measure between the color histograms or kernel density estimates of the target and candidate image. The most typically used similarity measures are the Bhattacharyya coefficient. There are assumptions which limited lighting condition due to color is very sensitive about illuminations. And the algorithm has weakness about inference of another object. In this paper we propose method that combined advantage of color distribution and depth. As apply robust error norm, problems are conquer. The method is useful for face tracking under the dynamic illumination. Also it voids an interference of another object.