Sign Language to Text Conversion using Deep Learning

Signs and gestures are a way of communication that, instead of spoken words, makes use of movements and gestures made by hands along with changes in facial expression and changes in bodily movements. There are many distinct sign languages throughout the world, just like there are many different spoken languages. Signs and gestures are primarily used by people who cannot hear and speak and use this to express their emotions and interact with each otherIn ancient times, gestures and signs were the only way of interacting with each other, since there were no spoken languages and was considered an obvious and natural way of communicating. However, as the time passed, the utilization of sign language started becoming more prevalent only for the people who are deaf or have any hearing impairments.Currently, over 80 million people in the world use gestures made by hands and symbols, which is a small part of the 780 crores people living on Earth as of 2020. These people depend greatly on sign languages to learn, access services and be a part of the communities. However, there is a lack of support in terms of providing basic services because of the fewer sign interpreters available.To overcome the challenges and issues created due to the dearth of information and knowledge relating to sign languages and to provide services that people deserve, there is a need to spread the use of sign languages among the general public. However, this involves a lot of effort and does not provide the results that are equivalent to the labour done. Detecting the correct signs in images can be a complex process and can include a variety of pre-processing techniques to be performed. In this project, various images containing the signs of the American-Sign-Language have been processed and utilised for implementing a CNN model to classify these images as their alphabetical equivalent.