Robust hand gesture recognition system using motion templates

Abstract

This paper presents a robust hand gesture analysis system. The approach uses the video analytic technique of motion templates rather than conventional gesture recognition algorithms. Also, it utilizes background modeling and skin pixel detection which further strengthens the approach by making it tolerant to background clutter and noise. In addition it reduces the false detections to a considerable extent. The system does not necessitate the user to wear any coloured caps or gloves for the hands. Encouraging results were obtained and it was found that the methodology is flexible and can be manipulated to suit gesture based interaction as per the requirements of a system. It can also be implemented as a standalone system.