The Application of Improved Camshift Algorithm in Hand Tracking

Camshift algorithm is often used in hand tracking. The traditional Camshift algorithm can realize hand tracking in simple condition, but it is easy to fail to track when similar color background interference and occlusion happens. In order to solve the problem, a kind of improved Camshift algorithm is proposed. Background subtraction algorithm based on mixed gaussian model is used to extract movement information of hand; then combine the color feature with movement feature by adopting an adaptive method to improve the algorithm. In the meantime, the Kalman filter algorithm is used to predict the location of hand, so the hand can be tracked accurately when it is covered. The experimental results show that improved Camshift algorithm can accurately track the hand when similar color background interference and occlusion happens, and improve the robustness of hand tracking.