Improving Accuracy of Hand Gesture Recognition using Recurrent Neural Networks

Abstract

In human-device communications, human gestures are crucial. Furthermore, hand-activated communication helps control without physical contact [1]. While the importance of hand gesture recognition techniques is rising, hand gesture recognition evidence has a low degree of reliability. To identify gestures, first and foremost, a wirelessly recognizable system is needed. Cameras, radar, and other options are available. Cameras, on the other hand, are impossible to use in environments with no sun, rain, or where personal privacy can not be violated. As a result, cameras used to be the chosen system, but due to different constraints, they now tend to use radar.