Sign language is a very different form of language for communication, which is important for a large group of people. The sign language has different signs which include hand shapes, motion profile, and position of the hand, face, and other body parts, each of which contributes to a new sign. Thus, it is a very complex task and virtual sign recognition is a complex research area in computer vision. There are many models this field, many of which are very efficient, which is achieved by deep learning approaches. In this, we review the vision-based proposed models of sign language recognition using machine learning approaches. While the overall trend of the proposed models indicates a significant improvement in recognition accuracy in sign language recognition, there are some challenges yet that need to be solved.