A new approach of hand tracking based on integrated optical flow analyse

In this paper, we focus on the hand tacking with a combinatorial restrained optical flow algorithm instead of the feature point. The movement region of hand is segmented and then searched for the target. A new template is used for hand target recognition on the segmented regions for detecting the ROI. Then optical flow is used to initialize step sizes in a stochastic optimization setting for hand tracking. For decrease the calculation cost, integrated additional information based on image sequence in an edge-based pose estimation computation is adopted as an accessorial instrument. The experimental results achieved show a more significative increase in the tracking speed than classical algorithms.