A Comprehensive Review of Real-Time Sign Language Interpretation Using Various Machine Learning Models

Real-time Sign Language interpretation is among the fastest-growing fields that can potentially minimize the communication gap between deaf people and the hearing population and prepare the ground for inclusivity and accessibility throughout social, educational, and professional contexts. The review discusses the development of technologies for sign languages interpretation in real time, having the main emphasis that have to be met such as high interpretation accuracy, low latency, and computational efficiency. This review methodology, which compared machine learning and deep learning algorithms and discussed different model architectures, analyzed the trends and emerging methodologies that have shaped the current state of the field and identified significant gaps in the literature. In particular, the tendency is such that the multimodality involving hand, face, and body movement participation is becoming increasingly more effective; the pre-trained models get adapted to perform sign language tasks. Most importantly, diversity in datasets promotes model robustness across languages and dialects. A bit further, practical problems are discussed with real-time interpretation deployment on mobile devices and other resource-limited platforms. The following review is designed to point out the likely best directions for future research and development that would generalize, through the advancement of real time sign language interpretation, to become an accessible and reliable technology for everyday use.