Bridging the Gap: Real-Time Telugu and Tamil sign Language Recognition using AI and computer vision

The main mode of communication for people who are deaf or hard of hearing is sign language. To communicate effectively, these individuals often rely on interpreters or professionals to convert their signs into spoken language. This creates a specific challenge for Tamil and Telugu-Speaking users, as there isn’t a specialized system for recognizing Tamil and Telugu sign language, forcing them to utilize more commonly used but unfamiliar languages. This dependence on interpreters can result in communication lag and potential misunderstandings. This project seeks to address the communication barrier by creating a real-time gesture recognition system utilizing the YOLOv5 algorithm and computer vision methods. In contrast to conventional models like CNNs or RNNs, YOLOv5 provides superior speed and accuracy via a pre-trained model that can adapt in real time. The system instantly converts identified gestures into Tamil and Telugu scripts, facilitating immediate and impactful communication. This initiative is the first of its kind dedicated to translating Tamil and Telugu sign languages. It employs a custom-trained dataset that includes 52 letters from Telugu, 247 letters from Tamil, 10 numerals, and 8 frequently used words. This strategy not only fosters inclusivity but also showcases the possibilities of AI-powered sign language systems for users of regional languages.