VSpSR: Explorable Super-Resolution via Variational Sparse Representation

Super-resolution (SR) is an ill-posed problem, which means that infinitely many high-resolution (HR) images can be degraded to the same low-resolution (LR) image. To study the one-to-many stochastic SR mapping, we implicitly represent the non-local self-similarity of natural images and develop a Var...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhou, Hangqi, Huang, Chao, Gao, Shangqi, Zhuang, Xiahai
Format: Conference Proceeding
Language:English
Subjects:
Computer vision
Conferences
Dictionaries
Neural networks
Space exploration
Superresolution
Visualization
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Super-resolution (SR) is an ill-posed problem, which means that infinitely many high-resolution (HR) images can be degraded to the same low-resolution (LR) image. To study the one-to-many stochastic SR mapping, we implicitly represent the non-local self-similarity of natural images and develop a Variational Sparse framework for Super-Resolution (VSpSR) via neural networks. Since every small patch of a HR image can be well approximated by the sparse representation of atoms in an over-complete dictionary, we design a two-branch module, i.e., VSpM, to explore the SR space. Concretely, one branch of VSpM extracts patch-level basis from the LR input, and the other branch infers pixel-wise variational distributions with respect to the sparse coefficients. By repeatedly sampling coefficients, we could obtain infinite sparse representations, and thus generate diverse HR images. According to the preliminary results of NTIRE 2021 challenge on learning SR space, our team ranks 7-th in terms of released scores.
ISSN:2160-7516
DOI:10.1109/CVPRW53098.2021.00047