A Method of Hand Gesture Segmentation and Tracking with Appearance Based on Probability Model

A novel probability model is presented using the probability-based method of moving information selection together with the skin color probability distribution image. After the division of the images, the moving information is obtained by the use of the statistical frame difference method. In order to locate the accurate location of the hand position, the center of the hand is found via moving value and skin color information. From the seed point, region growth method is proposed for the initiate of the hand region to obtain the size and position of a search window. The current size and location of the tracked hand are used to forecast and track the search window in the next video image. Then the segmentation and tracking of the dynamic hand are realized. Experiments show that this method is so robust to noise and complex backgrounds that it can track the moving object with the shape of hand changing. Additionally, tracking the hand accurately is feasible in a real-time way. Furthermore it can be used for the segmentation and the tracking of a moving object with special color.