Tamil Text to Sign Language Avatar System Enhancing Accessibility for the Hearing Impaired
R Alexander, R Ughan, Sai Kishor M D · 2025
Communication barriers between Indian Sign Language (ISL) users and Tamil speakers create significant challenges for individuals with hearing and speech impairments. To address this, we propose a two-way translation model utilizing a Flaskbased CNN-LSTM architecture for real-time ISL-to-Tamil and Tamil-to-ISL translation. The system is trained on a curated dataset comprising commonly used words in both languages, ensuring accurate and context-aware translations. Through rigorous testing, the model achieved an accuracy of 83%, demonstrating its reliability in facilitating seamless communication. This project serves as a crucial step toward inclusivity, with applications in language learning, real-time interactions, and accessibility technologies.