Human Tracking by Fast Mean Shift Mode Seeking

Abstract

Change detection by background subtraction is a common approach to detect moving foreground. The resulting difference image is usually thresholded to obtain objects based on pixel connectedness and resulting blob objects are subsequently tracked. This paper proposes a detection approach not requiring the binarization of the difference image. Local density maxima in the difference image - usually representing moving objects - are outlined by a fast non-parametric mean shift clustering procedure. Object tracking is carried out by updating and propagating cluster parameters over time using the mode seeking property of the mean shift procedure. For occluding targets, a fast procedure determining the object configuration maximizing image likelihood is presented. Detection and tracking results are demonstrated for a crowded scene and evaluation of the proposed tracking framework is presented.