Real-Time Efficient Detection in Vision Based Static Hand Gesture Recognition

The focus on Human-Computer Interaction (HCI) research is increasing day by day, due to the increasing requirement of intelligent input devices. Hand Gesture Recognition is a small sub-field but presents a significant number of applications and consumer products. Most researches target on the feasibility of recognition systems but give less weight to the device resources, so the cost and time. The time-consuming complicated algorithms' use is limited to special purpose devices such as expensive gaming consoles. The use of such systems in low cost embedded hardware in realtime circumstances is required, with the comfortability to use it. In this paper, we design an efficient real-time keyboard-like HCI using Static HGR. We have proposed and implemented new methods to reduce the time consumption while maintaining the high accuracy of 90% with scale and rotation invariance. Also, to maintain the comfort of use, we have eliminated complicated gestures and used only 11 gestures as input gesture set.