Blind Single Image Super-Resolution via Iterated Shared Prior Learning

Thomas Pinetz*, Erich Kobler, Thomas Pock, Alexander Effland

*Korrespondierende/r Autor/-in für diese Arbeit

Publikation: Beitrag in Buch/Bericht/KonferenzbandBeitrag in einem KonferenzbandBegutachtung

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.

Originalspracheenglisch
TitelPattern Recognition - 44th DAGM German Conference, DAGM GCPR 2022, Proceedings
Redakteure/-innenBjörn Andres, Florian Bernard, Daniel Cremers, Simone Frintrop, Bastian Goldlücke, Ivo Ihrke
Herausgeber (Verlag)Springer Science and Business Media Deutschland GmbH
Seiten151-165
Seitenumfang15
ISBN (Print)9783031167874
DOIs
PublikationsstatusVeröffentlicht - 2022
Veranstaltung44th DAGM German Conference on Pattern Recognition: DAGM GCPR 2022 - Konstanz, Deutschland
Dauer: 27 Sept. 202230 Sept. 2022

Publikationsreihe

NameLecture Notes in Computer Science
Band13485
ISSN (Print)0302-9743
ISSN (elektronisch)1611-3349

Konferenz

Konferenz44th DAGM German Conference on Pattern Recognition
KurztitelDAGM GCPR 2022
Land/GebietDeutschland
OrtKonstanz
Zeitraum27/09/2230/09/22

ASJC Scopus subject areas

  • Theoretische Informatik
  • Informatik (insg.)

Fingerprint

Untersuchen Sie die Forschungsthemen von „Blind Single Image Super-Resolution via Iterated Shared Prior Learning“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren