Interpretation of Hand Gestures Using Neural Networks:A Review

Since its emergence from the early 1980s, the field of Human Computer Interaction has moved on and advanced in many significant ways. It has opened up a world in which communication between human and computer has become easier and richer. Among the different modes of interaction, Gestures provide the most natural and convenient way of communication. Hence gesture recognition has been extensively researched and many systems have been implemented. However because of the complexity in understanding gestures, to date, we have not succeeded in devising a perfect gesture recognition system. Gesture recognition involves 4 stages out of which the final gesture recognition stage is the most complicated. Various approaches have been studied in recognizing gestures with varying degrees of success. In this paper, we discuss several studies which employ Artificial Neural Networks as classifiers. The review focuses on those systems that use the Feed Forward Networks as classifiers, mainly Multi-Layer Perceptron Networks, Back Propagation Networks and Radial Basis Function Networks. Keywords— Neural Networks, Human Computer Interaction, Back Propagation Neural Networks, Multi-Layer Perceptron, Radial Basis Function Networks, Gesture Recognition System