Video super-resolution is a critical task in medical imaging, particularly for radiological structures. It involves enhancing the resolution of low-quality video sequences to improve the visualization and analysis of medical images. Context-based learning in video super-resolution has become a promising technique for improving the resolution and quality of radiological structures in medical imaging. This technique leverages contextual information within video frames to improve the accuracy and visual fidelity of the super-resolved images. By considering spatial and temporal relationships, context-based learning methods aim to capture fine details and motion dynamics present in radiological structures. This survey aims to offer a complete overview of different learning techniques in video/image super-resolution.