Target tracking based on multiple feature and particle swarm optimization

Abstract

In this paper a new algorithm is proposed based on multiple feature and particle swarm optimization for target tracking. The multiple feature includes local binary pattern (LBP), phase congruent (PC) and gradient magnitude (GM). This method not only can make use of contrast invariance of phase congruent, but also can fully utilize the rotation invariance of local binary pattern. Compared with the traditional histograms, the proposed algorithm extracts effectively the edge and corner feature in the target region, which characterizes better and more robustly represents the target. The experiments show that the proposed method in this paper is more accurate and more efficient in tracking objective than the traditional algorithms.