Adaptive Ensemble Learning for Real Time Sign Language Recognition

This research provides a smoother way for detecting real-time sign language using adaptive ensemble learning. Still, 72 million people around the world rely on sign language recognition as their primary means of communication. There is a need for a solid and real-time sign language recognition system. When it comes to multilingual language recognition, current models still lack the time factor, efficiency, and scalability. To provide a desired solution, our proposed research work uses convolutional neural networks for attribute detection with the help of recurrent neural networks to identify real-time signs. The proposed recognition system resolves the problems in the existing models and focuses more on providing real-time results on a multilingual platform, including Indian Sign Language. The proposed model consists of various steps, like data preprocessing, ensemble learning techniques, and multimodal results generation capability. This work contributes to providing real-time sign language recognition on various platforms with different environmental conditions and possesses the capability of transforming communication into various sectors like finance, education, healthcare, smart industries, etc.