Tamil Alphabet Sign Language Recognition Using Mediapipe and Machine Learning

Tamil is an ancient and unique language with 247 characters, and it can be challenging for the deaf and mute community to communicate its alphabets through sign language to others, often leading to social exclusion. To bridge this communication gap, we propose a real-time Tamil sign language recognition system using the Mediapipe framework and various machine learning algorithms. Our study trained multiple models, including Support Vector Machine (SVM), Random Forest, XG-Boost, and Gradient Boosting, using a publicly available dataset specifically developed for Tamil letters and their corresponding signs. Among these models, the Random Forest classifier demonstrated the best performance, achieving an accuracy of 96.45%. This outcome provides a state-of-the-art, efficient real-time Tamil sign language translator, enhancing communication and promoting inclusivity within the Tamil community.