Background and skin colour independent hand region extraction and static gesture recognition

Hand extraction and gesture recognition has always been a challenging problem in its general form. In this paper, we consider a fixed set of standard gestures and a reasonably structured environment and develop three effective procedures for extracting hand from the image, two of which are for plain non-complex static background and one for complex static background making it independent of the skin and background colours. The second part is of recognizing the gesture and making it scale and rotation invariant. For hand extraction, the three basic concepts used are 1. Gaussian distribution, 2. K-Mean classification and 3. Simple background subtraction and consecutive frame subtraction to find the palm region in the complete image. In gesture recognition, we extracted some features like centre of hand region, no. of fingers and the distance between the fingers. Using these features, the gestures are classified into seven standard hand gestures.