Robust segmentation of hands in a cluttered environment for hand gesture recognition has remained a challenge in computer vision. In this work, a two-stage gesture recognition framework is proposed.