MSU Video Super-resolution Dataset

MSU Video Super Resolution Benchmark is a dataset for video super-resolution tasks. The dataset contains the most complex content in detail restoration tasks: faces, text, QR codes, license plates, textures without displayed patterns, small details. The videos include different types of motion and different types of degradation: bicubic interpolation (BI), Gaussian blur and downsampling (BD). The resolution of all input video sequences is 480×320.