Rakshit Kaushik

How can we accurately translate real-life Indian Sign language to text and text to Indian Sign Language?

Started May 16, 2022

Abstract or project description

One of the non-verbal communication techniques utilised in sign language is the hand gesture. It is mostly used by deaf and dumb individuals to communicate with other people or among themselves when they have hearing or speech issues. Many makers around the world have created various sign language systems, but they are neither flexible nor cost-effective for end users. Therefore, my goal in this project is to create software that can automatically identify sign language, enabling deaf and dumb people to interact with hearing people or normal people more successfully. The two emerging topics of study are pattern recognition and gesture recognition. Hand gestures play a crucial role in nonverbal communication and are essential to daily living. Computer recognition of sign language deals with translating real-life Indian sign language symbols to text in either English or Hindi. Sign gestures can be classified as static and dynamic. However static gesture recognition is simpler than dynamic gesture recognition, as it requires analysing a picture instead of a video, which is just a collection of multiple pictures. In this research, I will be working on how to best build my AI model, by working on the pre-processing and data augmentation, in order to have the most accurate data model, which can work on general data. I will also work on deciding which layers to use for my AI model in addition to the CNN to increase its accuracy. As the final deliverable, I hope to build a website where this model can be used by people to translate ISL pictures to text easily.