Static gesture segmentation technique based on improved Sobel operator

Abstract

The gesture recognition technology based on computer vision can provide a more natural way of human–computer interaction, and become a research hotspot in the field of gesture recognition. This study proposes an improved Sobel operator which is used for static hand contour extraction. The method is based on the noise existing in the picture, and a bilateral filtering algorithm is employed to denoise and preserve the edges to a large extent. Based on the colour space, the elliptical skin colour model is used to extract the skin colour domain, and the hand is separated according to the block domain feature. Finally, in order to solve the oblique indirection problem of the traditional Sobel operator, two oblique direction templates are added and the weights are reassigned. Then the improved Sobel operator is used to extract the gesture edges, and compared with the classical extraction edge operator Laplace and Canny extraction effects. The experimental results show that the proposed algorithm has high efficiency and accuracy in static gesture segmentation in a simple environment.