A visual-based research on static gesture recognition

Now vision-based gesture recognition plays an important role in the fields of image processing, pattern recognition and so on. Human being hands are highly variable organs and hand features are affected easily by various environmental factors. Considering the characteristics of hand gesture, in this paper we proposes an improved YCbCr color space method to segment gesture images, and extracts Fourier descriptors and Hu moment as recognition features. Finally, the Hausdorff distance is applied to recognize the gestures by the method of model matching. Experimental results show that the proposed method has higher operation and recognition rates.