Super Resolution is a process of transforming low-resolution video images into high-resolution. In this context, this work presents studies on different models based on deep learning that seek to solve the problem described. Throughout this work these selected models are described and intuitions behind their functioning are given. Furthermore, the positive and negative points of these different approaches are shown, presenting the results in the form of graphs, metrics, images and videos. These results are then analyzed and compared. To ensure clear understanding and explanation of the points raised, an in-depth study of some important concepts within the scope of deep learning are presented and exemplified.