Video super-resolution based on deep learning: a comprehensive survey

Video super-resolution (VSR) is reconstructing high-resolution videos from low resolution ones. Recently, the VSR methods based on deep neural networks have made great progress. However, there rarely systematical review these methods. In this survey, we comprehensively investigate 37 state-of-the-art learning. It well known that leverage of information contained in video frames important for super-resolution. Thus propose a taxonomy and classify into seven sub-categories according to ways utilizing inter-frame information. Moreover, descriptions architecture design implementation details are also included. Finally, summarize compare performance representative some benchmark datasets. We discuss applications, challenges, which need be further addressed by researchers community VSR. To best our knowledge, work first systematic tasks, it expected make contribution development recent studies area potentially deepen understanding techniques