Key frame-based video super-resolution using bi-directional overlapped block motion compensation and trained dictionary

Abstract

This paper presents a video super-resolution algorithm to interpolate an arbitrary frame in a low resolution video sequence from sparsely existing high resolution key pictures. Firstly, hierarchical motion estimation is performed between the input and the key low resolution frames on a patch basis. If motion-compensated error is small, an input LR patch is super-resolved by bi-directional overlapped block motion compensation. Otherwise, the input patch is spatially super-resolved using the dictionary which is learned from the key low resolution and its corresponding high resolution pictures beforehand. Finally, possible blocking artifacts between a motion-compensated patch and a spatially super-resolved patch are concealed using a specific de-blocking process. Experimental results show that the proposed algorithm provides significantly better subjective visual quality as well as objective quality than previous interpolation algorithms.