Language disparities have always existed in human civilization. In their community, deaf individuals use sign languages to communicate with one another, but it can be difficult to do so when interacting with the wider population. The project comprises a real-time sign language recognition system that enables hearing-impaired individuals and persons who do not know sign language to effortlessly communicate. The most used sign language in South Asian nations is Indian sign language. Convolution neural networks, specifically Tensor Flow object detection, are utilized to assess and train the model on the images that are used. The transfer training approach is used to train the deep learning model with various hand movements. Loss is anticipated to be decreased.