Sign language is the mode of communication for such people who are not blessed with the gift of hearing and speech. It involves the use of hands and facial expressions, Understanding sign language is difficult for people with no prior knowledge. This leads to a communication gap for such people. An automated sign language identification system can aid in reducing this gap. Here, a system is presented to automatically identify sign language. Experiments were performed on the publicly available sign language MNIST dataset. The highest accuracy of 99.98% was obtained in this study with a convolutional neural network-based approach.