Human Interface Intelligent Gesture Recognition System for Hearing Impaired Communities

Abstract

The development of advanced human-computer interaction technologies has led to a significant improvement in communication and collaboration in various industries. However, there are still many individuals with special needs who face communication barriers in the workplace, which can hinder their ability to perform at their best. This challenge highlights the need for innovative solutions that can bridge the communication gap and support effective collaboration between all individuals. To address this challenge, this research proposes the development of a human interface intelligent gesture recognition system that utilizes cuttingedge technology to recognize and interpret a range of gestures with lightning-fast speed and accuracy. The system generates equivalent text in real-time, allowing for seamless communication between individuals with different communication abilities. It includes the use of an innovative methodology to sort out certain limitations of gesture recognition systems, such as determining higher gesture value frames from selected images for both alphabet and sign word identification. First, a Gradient-based Machine Learning (GML) classifier is proposed in a hybrid pre-trained 3D-CNN framework to reduce the probability chance of selecting gesture similarity index alphabetic image frames. An innovative method, called Probability of Predicting the Concurrent Gestures (PPCG), is introduced to predict the gestures of the necessary alphabet from the observed gesture motion. Second, a Convolutional Deep Visual Geometry Group (CDVGG-16) classifier is used in a hybrid pre-trained 3D-CNN framework to improve the probability selection of the best gesture frames for sign word identification newline