Research On Tracking Algorithm Of Visual Face

Human face tracking (HFT) is one of several technologies useful in vision-based interaction (VBI). Robust face tracking is a prerequisite for face analysis and recognition. However, it is a challenging issue in computer vision that tacking objects efficiently and robustly in complex environment.In this context, a novel algorithm is proposed for scale adaptive face tracking. Mean-Shift process is applied to Particle Filter (PF) framework, where PF is optimized by Mean-Shift procedure due to its property of accurate tracking and fast convergence. And the accuracy position determination makes window size computation possible in the proposed method. On the other hand the algorithm maintains multi-hypothesis tracking, so it is still robust. Additionally, an effect mechanism is proposed for checking tracking result, which provides useful information for next frame. In this paper, we have described and evaluated the performance of the proposed algorithm. The experimental results verify its efficiency and robustness for face tracking.