A novel system of real time hand tracking and gesture recognition

This paper presents a novel system of hand tracking and gesture recognition in real time domain. In this paper, SIFT algorithm is used to generate image features which are invariant to image translation, scaling, rotation and partially invariant to illumination changes. The extracted key points are matched with saved key points and the image is recognized. We carried out experimentations on MATLAB platform and have successfully identified a number of gestures. To prove the versatility of our system, we have used different light conditions in real time. The obtained results are fast and encouraging.