Educational Learning-Based Sign Language System Using Machine Learning

Abstract

This study proposes an innovative approach to multicultural education by integrating Indian Sign Language (ISL) and American Sign Language (ASL) through Machine Learning (ML) techniques. By collecting and preprocessing high-quality video data of ISL and ASL, we aim to develop ML models capable of recognizing and generating signs in both languages. Through bidirectional transfer learning and cross-language representation learning, we seek to enhance the learning experience and address common challenges in sign language acquisition. Additionally, personalized learning environments and culturally sensitive design, informed by collaboration with Deaf communities in India and America, ensure inclusivity and accuracy. Evaluation metrics and ethical considerations are integrated into the development process to promote responsible implementation and continuous improvement. Ultimately, this project aims to lay the groundwork for advancing multilingual sign language education globally. By employing advanced ML techniques, this study aims to bridge the gap between Indian Sign Language (ISL) and American Sign Language (ASL) education, fostering inclusivity and accessibility in learning environments. Through meticulous data collection, preprocessing, and collaborative development processes, our approach emphasizes accuracy, cultural sensitivity, and personalized learning experiences. By engaging with Deaf communities in both India and America, we ensure the authenticity and relevance of our platform. Evaluation metrics and ethical considerations are prioritized to uphold privacy, consent, and fairness principles. By establishing a robust foundation for multilingual sign language education, this project contributes to broader discussions on leveraging ML for enhancing accessibility and inclusivity in education systems worldwide. Keywords- Hand Gesture, Sign language Recognition, OpenCV, Media-pipe, tensorflow.