Literature review on dynamic hand gesture recognition

Abstract

Sign language is a method for the dumb and deaf society to connect with the exterior world. The dumb and deaf people find it very difficult to express their feelings to the normal people, since the normal people have the insufficiency knowledge of the sign language used by the mute and deaf people. Thus the development of a system to fill the communication gap in between mute/deaf and the normal people who are well able to speak and hear would be of great utility to the society. A sign language recognition model involves of accurate and effective tool to translate sign language into text/speech. Gesture recognition identifies a significant expressions of a man-made gesture by hands. Gesture identification from video sequences has significant deviations in the field of behavioral understanding and computer vision. This will provides a platform for conversion of predictable gestures in sentences that are cumulative of words used in an approved sequence. With the advance of deep learning, artificial intelligence has completed significant achievement in modern growths in gesture recognition technology for videos based on deep learning. As compare to others gesture recognition methods, dynamic hand gesture recognition using deep learning to progress the recognition accuracy. This paper presents a research gap survey on various hand gesture recognition methods. Hand gestures are categorized in two types-1) Static hand gestures and 2) Dynamic hand gestures. The first type as static hand gesture recognition is the fingerspelling level detection of alphabets, words and numbers. The second type as dynamic hand gesture recognition are used for identification of sign language sentences. Sign language recognition systems using gesture play a significant part in Human-computer interaction (HCI) to establish communication between dumb/deaf societies and normal people. In this paper, a literature review on hand gesture recognition methods are presented.