File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Learning MRI artefact removal with unpaired data

TitleLearning MRI artefact removal with unpaired data
Authors
Issue Date2021
Citation
Nature Machine Intelligence, 2021, v. 3, n. 1, p. 60-67 How to Cite?
AbstractRetrospective artefact correction (RAC) improves image quality post acquisition and enhances image usability. Recent machine-learning-driven techniques for RAC are predominantly based on supervised learning, so practical utility can be limited as data with paired artefact-free and artefact-corrupted images are typically insufficient or even non-existent. Here we show that unwanted image artefacts can be disentangled and removed from an image via an RAC neural network learned with unpaired data. This implies that our method does not require matching artefact-corrupted data to be either collected via acquisition or generated via simulation. Experimental results demonstrate that our method is remarkably effective in removing artefacts and retaining anatomical details in images with different contrasts.
Persistent Identifierhttp://hdl.handle.net/10722/325507
ISI Accession Number ID
Errata

 

DC FieldValueLanguage
dc.contributor.authorLiu, Siyuan-
dc.contributor.authorThung, Kim Han-
dc.contributor.authorQu, Liangqiong-
dc.contributor.authorLin, Weili-
dc.contributor.authorShen, Dinggang-
dc.contributor.authorYap, Pew Thian-
dc.date.accessioned2023-02-27T07:33:51Z-
dc.date.available2023-02-27T07:33:51Z-
dc.date.issued2021-
dc.identifier.citationNature Machine Intelligence, 2021, v. 3, n. 1, p. 60-67-
dc.identifier.urihttp://hdl.handle.net/10722/325507-
dc.description.abstractRetrospective artefact correction (RAC) improves image quality post acquisition and enhances image usability. Recent machine-learning-driven techniques for RAC are predominantly based on supervised learning, so practical utility can be limited as data with paired artefact-free and artefact-corrupted images are typically insufficient or even non-existent. Here we show that unwanted image artefacts can be disentangled and removed from an image via an RAC neural network learned with unpaired data. This implies that our method does not require matching artefact-corrupted data to be either collected via acquisition or generated via simulation. Experimental results demonstrate that our method is remarkably effective in removing artefacts and retaining anatomical details in images with different contrasts.-
dc.languageeng-
dc.relation.ispartofNature Machine Intelligence-
dc.titleLearning MRI artefact removal with unpaired data-
dc.typeArticle-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1038/s42256-020-00270-2-
dc.identifier.scopuseid_2-s2.0-85099606902-
dc.identifier.volume3-
dc.identifier.issue1-
dc.identifier.spage60-
dc.identifier.epage67-
dc.identifier.eissn2522-5839-
dc.identifier.isiWOS:000612287900015-
dc.relation.erratumdoi:10.1038/s42256-021-00300-7-
dc.relation.erratumeid:eid_2-s2.0-85099758303-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats