Paper Title
AI-BasedReal-TimeTwo-WaySignLanguage Translator for Hearing and Speech Impaired Communication

Abstract
The successful communication between people with hearing or speech impairment and the rest of the society remains a major area of concern in the society primarily due to the lack of common understanding on sign language. This paper proposes a bidirectional real-time sign language interpretation system as an artificial intelligence-based system that can eliminate this communication barrier. The system accepts or captures the hand gestures and, in turn, the spoken input is translated to given sign gestures. It uses the computer vision techniques by OpenCV to handle live video links with the help of foreground segmentation resting on the concept of applying Gaussian Mixture Models, and hand regions. Gesture recognition is done through a Deep Neural Network, where gestures identified as such are translated to a textual message that is then translated into speech through Google Text-to-Speech (gTTS). On the other side, speech recognition is combined with speech natural language processing to produce suitable signs representations. The suggested solution is affordable, readily available, and dependable and has great potential to increase inclusivity in the areas of education, health care, workplace, and just regular social interactions. Keywords - Sign Language Translation, Computer Vision, Deep Neural Networks, Speech Recognition, Assistive Technology