DYNAMIC HAND GESTURE RECOGNITION USING SPATIAL-REMPORAL FETURES

Abstract

Hand gesture recognition has been studied for a long time. However it still is a challenge field. Furthermore, hand gesture recognition as a natural way to control devices in smart-house such as television, light, fan, camera, door often require high accuracy. This paper proposes a simple and effective technique to recognize a pre-defined hand gesture set. The defined dynamic gestures are represented in both hand shape changing during temporal dimension and direction of hand movements. In the proposed technique, we analyze Spatial - Temporal features that includes characteristics of a hand gesture such as cycle pattern, different in length, non-synchronization phase, After that we evaluate the proposed technique in term of recognition rate and computational time.