Communication is key to share and express ideas and thoughts. And not every person has the natural ability to communicate. Mute people communicate with each other and with rest of the people through sign language. Sign language is the way of communication using gestures with hands where each gesture corresponds to particular letter in an alphabet or a word. But to use this every person must have the knowledge of it. But in the real world, it is not the case. With the advancement in the field of deep learning it is possible to solve this problem with low cost. So, this project aims at making it easier for mute people to communicate to the rest of the world. In this project, image of the hand gesture made by the person is captured and is translated to its corresponding meaning in the American sign language system. Equipment such as hand gloves sensor is one of the methods used for sign language detection. But it is expensive and not all of the people in need of it would be able to afford it. A deep learning algorithm called convolutional neural network is used for identifying the gestures and giving the actual letter or word.