Multiple Features Human Face Tracking Based on Particle Filter
Haitao Yao, Fuxi Zhu · 2009
In this paper, a multiple features face tracking algorithm based on the particle filter is proposed. Since the particle filter can effectively combine multiple face features information which represents different characteristics, and supply robustness in different environments, we combine the robustness and invariance to rotation and translation of color histogram central moment and the accuracy and the less computation complexity of 2D RCWF in particle filter framework to propose a new human face tracking algorithm. Experimental results demonstrate the efficiency and effectiveness of the algorithm and show a more robust face tracking performance compared with methods based on single feature.