Hand Gesture Contour Tracking Based on Skin Color Probability and State Estimation Model

Abstract

Considering the deficiency of accurate hand gesture contour inaccessible and inefficiency in complex dynamic background in existing methods of hand gesture tracking, a two dimensional skin color probability forecast method is proposed. Based on this, a hand gesture segmentation method of multi-mode and a hand gesture tracking method of state estimation are extended. When hand gesture is segmented, to locate the accurate hand gesture position, this paper combines the Skin Color Probability distribution with the statistical motion information of image blocking. Then the hand region is initiated by the region growth method and the hand gesture segmentation is realized. When hand gesture is tracked, the pixel’s state model is built to estimate the state of pixels after watershed computation . Then the current blocking frame is adaptive threshold segmented and the hand gesture tracking is realized. Experiments show that this method has a strong anti-noise ability in complex background. In addition, it has a better application effect in segment and tracking the hand gesture contour accurately in a real-time way.