Seeing the Unheard: Lip Reading with Deep Learning

Communication can be difficult for people who are deaf and dumb. These people might face difficulties in understanding the spoken language in their everyday social lives. Traditional methods do not find an effective way to eliminate the communication barrier, and it also takes a long time to learn sign language for both the communicator and listener. There are chances of miscommunication, which can lead to misconceptions and problems. Therefore, a modern solution is required to eliminate this gap for such people, as technology is leading the way to solve worldwide problems and giving promising solutions. This paper puts forward a solution to this confrontation using deep-learning methods and convolutional neural networks with an accuracy of 93%. The whole model is trained on the GRID dataset containing 1000 different videos, wherein different sentences and phrases are spoken by the speaker.