Vad-Anuvad: Speech to Sign Language & Sign Language to Speech

A thorough explanation of a speech-to-sign language and sign language-to-speech converter system is provided in this work. Detecting and recognizing sign language, translating it to text, and then translating that text to speech are all part of the sign language-to-speech system. Our smartphone application uses the Flutter API for text-to-speech translation and YOLOv5 (You Only Look Once) as its object detection model to recognize sign language. When it came to object detection, YOLO was first evaluated with several iterations. Based on the outcomes of the model training procedure, YOLOv5 produced the most accurate findings. We created a speech-to-text model and compared its accuracy to the Flutter API for the purpose of translating speech to sign language. The Flutter API was selected as the more realistic model to translate speech to text. We found two ways to translate text to sign language: one way is to show pictures of signs, and the other is to use a SiGMl player based on HamNoSys that creates 3D models to display sign language. Our main objective was to develop an end-to-end integrated mobile application that would function as a barrier to entry for users of any technical skill level by utilizing our machine learning model to execute detections using the power of a local computer.