Feature point tracking based on RLS and MAP filter

Human motion tracking is crucial for many important applications. In this paper we propose an approach to human motion tracking from monocular image sequences. First, a system is developed for solving the occlusion problems. The system is based on recursive least square (RLS) and genetic algorithm (GA) that introduced a new way to eliminate occlusion. Then, in order to reduce the noise of position coordinates, the maximum a posteriori (MAP) estimator is jointed into the system. The tracking capability of proposed algorithm is proved. Experimental results on image sequences of different human motion, including walking and running, demonstrate the feasibility of the proposed approach.