Sign Language Detection Using Deep Learning and Tensorflow

Developing a hand sign detector which can recognize various hand signs made by the people who are unable to speak or facing difficulty to speak. The making of this project involves the use of mediapipe, which is an open-source framework developed to perform computer vision inferences. It also uses deep learning to classify images with pre-trained networks using Tensorflow. The mission of the proposed project is to use the hand signs in front of any device with built-in camera which can show what the hand sign’s actually means according to American Sign Language. It will detect hand points using cvzone and mediapipe, will use Tensorflow to identify the signs, and will give real-time output to the user; the proposed paper uses deep learning on self-generated input to classify data and will analyze output against the accuracy and time.