DYNAMIC HAND GESTURE RECOGNITION WITH RESPECT TO SIGN LANGUAGE

摘要

The focus of this research is to recognize ISL alphabets and sentences from video database and in real time . It is done by extracting different hand gesture features . ISL database accurately recognize ISL signs for both static and dynamic gestures. The palm orientation detection gave correct results to different hand poses. Also algorithms robustness is verified in varying background and all five ISL sentences with nine gestures are correctly recognize except skin colour background. From 44 Karpallavi alphabet data 41 static hand gestures are correctly recognized. Also from video database 10 Marathi sentence of hand gestures are correctly recognized and interpreted in audio form. The algorithm is deployed on embedded video image processing platform and able to recognize ISL number system and words. The algorithm is groomed up intelligently to tell no gesture, wrong gesture and insufficient light conditions. Also it is capable to add unknoen hand gesture on fly. The system will greatly benefit physically challenged deaf and dumb persons.