Maadhyam: Sign Language Translator (ASL)

Abstract

There are several million deaf and dumb people around the world and their only means of communication is sign language. They are usually not able to receive education and even some of them are unemployed due to lack of communication. This often leads to mental health issues among them. Deaf and dumb people rely on sign language interpreters for communications because other people do not prefer learning their language. However, our research aims to fulfil this gap, so that other people can understand sign and gestures without any human interpreter. It is a web application based on neural networks on which user can provide any random gesture and the project will automatically identify what gesture it is. We have currently trained our model for ASL (American Sign Language) letters only i.e. from A to Z. We have used CNN (Convolutional Neural Network) for training and recognizing gestures.