Hand Gesture Recognition with Skin Detection and Deep Learning Method

Abstract

Gesture recognition, although has been exploring for many years, is still a challenging problem. Complex background, camera angles and illumination conditions make the problem more difficult. Thus, this paper presents a fast and robust method for hand gesture recognition based on RGB video. First we detect the skin based on their color. Then we extract the contour and segment the hand region. Finally we recognize the gesture. The results of experiment demonstrate that the proposed method are efficient to recognize gesture with a higher accuracy than the state of the art.