A particle filter tracking method based on human visual system characters

On the basis of simulating the human eye's characters,in the process of the object tracking and modifying the particle filter framework,a particle filter tracking method based on human visual system characters is proposed.This method adopts the multi-cue fusion and utilizes the combination between the grey histogram and the motion region area as the criteria of the similarity of each particle,which improves the sensitivity about the region where the motion takes place.Particles are filtrated in the update stage and only the ones with comparative larger weight are picked out to judge the final tracking result,which guarantees the tracking precision.Experimental results show this tracking method is not only feasible but also with good performance.