Smart AI-based Indian Sign Language Translator for Bidirectional Communication Using Machine Learning

Communication barriers between people with speech impairments and the hearing population limit access to education, employment, healthcare, and public services. This paper presents a smart AI-based Indian Sign Language (ISL) Translator that is capable of supporting bidirectional interaction: Sign-to-Text and Text-to-Sign conversion. For Sign-to-Text, hand gestures captured via webcam are preprocessed and then classified using a Support Vector Machine (SVM) trained on a custom ISL dataset of static alphabet and digit signs. For Text-to-Sign, user-entered English text is tokenized and mapped to corresponding animated ISL videos from a curated Kaggle dataset, generating continuous sign sequences for multi-word input. The prototype achieves real-time performance with 95.8% accuracy in static gesture recognition and produces intelligible animated sequences for common phrases. The proposed system can function like an assistive interface to bridge communication gaps between Deaf and Hard of Hearing (DHH) users and hearing individuals in real-world environments.