Deep Neural Networks for Image-Based Indian Sign Language Recognition: A Comprehensive Review and Practical Analysis

Addressing the communication barriers faced by individuals with vocal and hearing disabilities, sign language serves as a visual means of expression. However, the limited understanding of sign language among the general population creates significant challenges in various settings, such as banks, airports, and supermarkets. To overcome this, a sign language recognition (SLR) system is essential. This study focuses on developing a real-time word-level sign language recognition system that translates sign language into text, with a specific emphasis on Indian sign language (ISL) to cater to the needs of the deaf and hard-of-hearing community in India. Our research presents an ISL-based sign language recognition system, enabling users to capture hand movement images through a web camera, with the system accurately predicting and displaying the corresponding sign name. The captured images undergo several processing stages, incorporating computer vision techniques like grayscale conversion, dilatation, and masking. To achieve robust recognition, a convolutional neural network (CNN) is trained on the collected dataset, demonstrating exceptional performance with a 99% accuracy rate. This research contributes to the advancement of assistive technologies by providing an effective solution for real-time ISL recognition, promoting inclusive communication and accessibility for individuals with hearing and vocal impairments.