Multiple Non-Rigid Object Tracking Using Fast Particle Swarm Optimization

Visual surveillance in crowded scenes, especially for humans, has recently been one of the most active research topics in machine vision because of its applications such as deter and response to crime, suspi-cious activities, terrorism or human behavior recogni-tion. Three of the most important problems in multiple human tracking are the occlusion problem, the non-rigid object tracking and run time. In this paper, we use particle swarm optimization (PSO) as a tracker, incre-mental subspace learning to introduce a likelihood func-tion for non-rigid objects and some other mathematical equations for multiple non-rigid object tracking. This paper modified PSO to improve the run time and the computational rate. Experimental results on simulated random data and several real video sequences from different conditions have shown the effectiveness of our approach