Particle filter tracking based on color and SIFT features

In the traditional particle filter tracking system, the weight of each particle is determined only by Bhattacharyya coefficient of two corresponding color histogram, which may easily lead to error tracking when the object and background have similar color distribution. In this paper, a novel particle filter algorithm is proposed in which the weight of particle is determined by both the color cue and local scale invariant feature transform (SIFT) features. The particle weight is calculated firstly by color similarity measurement and then updated according to the distribution of SIFT matches. The tracking window is designed to change size efficiently according to the former and current color and SIFT features. Experimental results show that the proposed method can effectively improve the tracking precision especially when the object is scale changing or in the clutter background of similar colors.