Intelligent visual object tracking with particle filter based on Modified Grey Wolf Optimizer

In a visual object tracking technology the particle filter (PF) is frequently used. The main drawback of the particle filter is that a large quantity of particles is required. This paper objectives to propose an evolutionary particle filter based upon Modified Grey Wolf Optimizer (MGWO) which will overcome the impoverishment of the sample problem in the regular PF. For this, firstly a new variant of GWO named as Modified Grey Wolf Optimizer (MGWO) is proposed. This variant works an active trigonometric sine truncated function for confirming the enhanced exploitation and exploration properties. Secondly, the MGWO algorithm is embedded in the PF structure. Before the resampling, by using MGWO, the particles in the PF are optimized. Accordingly, the more significant particles can be expanded, and the particles can estimate the actual state of the target object more precisely. Performance of proposed Modified GWO based particle filter (MGWO-PF) is evaluated on standard visual tracking benchmark databases. Also, the MGWO-PF tracker is compared with the Particle filter (PF), Particle swarm optimization based particle filter (PSO-PF), Firefly algorithm based particle filter (FAPF) and Spider monkey optimization assisted particle filter (SMO-PF). We show that visual object tracking using MGWO-PF provides more reliable and efficient tracking results than other compared methods.