Around 1 million to 2.7 million Indians use Indian Sign Language to communicate; this figure clearly states how important it is to augment a strong and dependable system for easy communication. Sign language is mostly learned by the deaf and dumb and is most likely unknown to others; as a result, communication becomes difficult. With the advancement of time, various approaches have emerged for smooth communication using sign language. The most common technique for interpretation involves using image processing algorithms to draw out features from coordinated motions, then applying convolutional neural networks (CNN) to master these characteristics and improve their functionality. An algorithm that can identify and predict objects in one forward pass is You Only Look Once (YOLO). In this study specifically Yolov7 i.e., the latest model of the Yolo series was used. The development of a system that enables individuals to utilise sign language independently can significantly aid them in being independent and ignite the confidence to exhibit their individuality to the public fearlessly. Therefore, it is crucial to create a hand gesture recognition system that can recognise hand signs in a developing nation like India with utmost accuracy. Our study dealt with developing a static hand gesture recognition algorithm to detect Indian Sign Language gestures used to communicate in our day-to-day lives. The machine learning model built was able to provide a perfect precision-recall graph, with 99% precision and recall.
Indian Sign Language, Convolutional Neural Networks, You Only Look Once
Sejal Vasan , Tushar Bhutani , Amita Goel , Nidhi Sengar , Vasudha Bahl "Real Time Recognition of Sign Language Using Convolutional Neural Network" Iconic Research And Engineering Journals Volume 6 Issue 9 2023 Page 29-36
Sejal Vasan , Tushar Bhutani , Amita Goel , Nidhi Sengar , Vasudha Bahl "Real Time Recognition of Sign Language Using Convolutional Neural Network" Iconic Research And Engineering Journals, 6(9)