Show simple item record

dc.contributor.authorKe, Rihuanen
dc.contributor.authorSchonlieb, Carola-Bibianeen
dc.date.accessioned2021-04-26T23:31:14Z
dc.date.available2021-04-26T23:31:14Z
dc.date.issued2021-04-05en
dc.identifier.issn0162-8828
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/321616
dc.description.abstractStandard supervised learning frameworks for image restoration require a set of noisy measurement and clean image pairs for which a distance between the output of the restoration model and the ground truth images is minimized. The ground truth images, however, are often unavailable or very expensive to acquire in real-world applications. We circumvent this problem by proposing a class of structured denoisers that can be decomposed as the sum of a nonlinear image-dependent mapping, a linear noise-dependent term and a small residual term. We show that these denoisers can be trained with only noisy images under the condition that the noise has zero mean and known variance. The exact distribution of the noise, however, is not assumed to be known. We show the superiority of our approach for image denoising, and demonstrate its extension to solving other restoration problems such as blind deblurring where the ground truth is not available. Our method outperforms some recent unsupervised and self-supervised deep denoising models that do not require clean images for their training. For blind deblurring problems, the method, using only one noisy and blurry observation per image, reaches a quality not far away from its fully supervised counterparts on a benchmark dataset.
dc.format.mediumPrint-Electronicen
dc.languageengen
dc.rightsAll rights reserved
dc.rights.uri
dc.titleUnsupervised Image Restoration Using Partially Linear Denoisers.en
dc.typeArticle
prism.publicationDate2021en
prism.publicationNameIEEE transactions on pattern analysis and machine intelligenceen
prism.volumePPen
dc.identifier.doi10.17863/CAM.68734
dcterms.dateAccepted2021-03-29en
rioxxterms.versionofrecord10.1109/tpami.2021.3070382en
rioxxterms.versionAM
rioxxterms.licenseref.urihttp://www.rioxx.net/licenses/all-rights-reserveden
rioxxterms.licenseref.startdate2021-04-05en
dc.identifier.eissn1939-3539
rioxxterms.typeJournal Article/Reviewen
pubs.funder-project-idEPSRC (EP/N014588/1)
pubs.funder-project-idEPSRC (EP/S026045/1)
pubs.funder-project-idEPSRC (EP/T003553/1)
pubs.funder-project-idEuropean Commission Horizon 2020 (H2020) Marie Sk?odowska-Curie actions (691070)
pubs.funder-project-idAlan Turing Institute (unknown)
pubs.funder-project-idEuropean Commission Horizon 2020 (H2020) Marie Sk?odowska-Curie actions (777826)
pubs.funder-project-idLeverhulme Trust (RPG-2018-121)
pubs.funder-project-idLeverhulme Trust (PLP-2017-275)
pubs.funder-project-idAlan Turing Institute (Unknown)
pubs.funder-project-idEPSRC (EP/T017961/1)
pubs.funder-project-idRoyal Society (RSWF\R3\193016)


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record