Blind Single Image Super-Resolution via Iterated Shared Prior Learning

Thomas Pinetz, Erich Kobler, Thomas Pock, Alexander Effland

Research output: Chapter in Book or Conference ProceedingsConference Proceedings without Presentationpeer-review

Abstract

In this paper, we adapt shared prior learning for blind single image super-resolution (SISR). From a variational perspective, we are aiming at minimizing an energy functional consisting of a learned data fidelity term and a data-driven prior, where the learnable parameters are computed in a mean-field optimal control problem. In the associated loss functional, we combine a supervised loss evaluated on synthesized observations and an unsupervised Wasserstein loss for real observations, in which local statistics of images with different resolutions are compared. In shared prior learning, only the parameters of the prior are shared among both loss functions. The kernel estimate is updated iteratively after each step of shared prior learning. In numerous numerical experiments, we achieve state-of-the-art results for blind SISR with a low number of learnable parameters and small training sets to account for real applications.
Original languageEnglish
Title of host publicationPattern Recognition. 44th DAGM German Conference, DAGM GCPR 2022,Konstanz, Germany, September 27-30, 2022, Proceedings
EditorsBjörn Andres, Florian Bernhard, Daniel Cremers, Simone Frintrop, Bastian Goldlücke, Ivo Ihrke
PublisherSpringer
Pages151-165
Number of pages15
ISBN (Print)978-3-031-16788-1
DOIs
Publication statusPublished - 2022

Research Field

  • High-Performance Vision Systems

Keywords

  • Explainable AI
  • Machine Learning

Fingerprint

Dive into the research topics of 'Blind Single Image Super-Resolution via Iterated Shared Prior Learning'. Together they form a unique fingerprint.

Cite this