Visual tracking using quantum-behaved particle swarm optimization

Abstract

Visual tracking is one of the most important applications in computer vision. Since the tracking process can be formed as a dynamic optimization problem. PSO, an effective algorithm to solve optimization problem, has been used in tracking widely. However, it has been proved that the traditional PSO is easy to converge to local optimum. In this paper, we adopt quantum-behaved particle swarm optimization (QPSO) for visual tracking. QPSO has better global convergence compared with the PSO, and can overcome the shortcomings of PSO algorithm. In order to achieve better tracking performance, we improve the traditional tracking framework based on PSO and propose a sequential QPSO based tracking algorithm in this paper. We conduct numerous experiments, and the results have shown the effectiveness of our method, even when the object undergoes abrupt motion or large changes in illumination, scale and appearance.