Real time face tracking by genetic particle filter

There are a great variety of human faces tracking methods based on particle filter. However, most tracking algorithms, so far, are unable to meet the demands for both precise and fast tracking. A real-time algorithm, based on genetic particle filter (GPF) for human faces tracking is presented in this paper. The crossover and mutation operations in evolutionary computation are introduced into PF to make samples move towards regions with large value of posterior density function (PDF). Experiments results show that GPF presents improvements over the PF techniques regarding to robustness, accuracy and flexibility in dynamic environment. Meanwhile, GPF, which needs fewer samples, improve the speed of tracking.