A Comprehensive Review on Real-Time Sign Language Detection for Deaf and Dumb People

Sign language is a critical mode of communication for individuals with hearing and speaking impairments. This research aims to create a precise and efficient model capable of accurately recognizing and translating manual sign gestures into text or speech. The system is trained using a combination of supervised and self-supervised learning approaches, where both labeled and unlabeled data are utilized to improve performance. There are several language models and datasets are available for languages like English but there is limited work has been done on the Marathi language.