3D Hand Gesture Detection Using Deep Learning Algorithms

3D hand gesture recognition has significantly improved human-computer interactions due to the advancement in deep learning, depth sensors, and artificial intelligence. These technologies have enhanced the development of convenient, compatible hand recognition interfaces, thereby making it possible for people with disabilities to interact with others efficiently. The aim of this paper is to evaluate and compare 3D hand gestures using several metrics such as execution time, training, testing, sensitivity, specificity, positive and negative predictive value, and likelihood. Hence, this paper proposes the use of convolutional neural network (CNN) algorithms to recognize 3D hand gestures. Two people participated in the whole experiment, and a total of twelve 3D hand motions were recorded within a long distance. Experimental results show that the first participant in the single experiment and the second participant in the combined experiment have the best values in most parameters.