Adaptive diffusion priors for accelerated MRI reconstruction

Simple item page
buir.contributor.authorGüngör, Alper
buir.contributor.authorDar, Salman Ul Hassan
buir.contributor.authorÖztürk, Şaban
buir.contributor.authorKorkmaz, Yılmaz
buir.contributor.authorBedel, Hasan Atakan
buir.contributor.authorElmas, Gökberk
buir.contributor.authorÖzbey, Muzaffer
buir.contributor.authorÇukur, Tolga
buir.contributor.orcidGüngör, Alper|0000-0002-3043-9124
buir.contributor.orcidÇukur, Tolga|0000-0002-2296-851X
dc.citation.epage102872-16en_US
dc.citation.spage102872-1
dc.citation.volumeNumber88
dc.contributor.authorGüngör, Alper
dc.contributor.authorDar, Salman Ul Hassan
dc.contributor.authorÖztürk, Şaban
dc.contributor.authorKorkmaz, Yılmaz
dc.contributor.authorBedel, Hasan Atakan
dc.contributor.authorElmas, Gökberk
dc.contributor.authorÖzbey, Muzaffer
dc.contributor.authorÇukur, Tolga
dc.date.accessioned2024-03-17T08:35:13Z
dc.date.available2024-03-17T08:35:13Z
dc.date.issued2023-07-20
dc.departmentDepartment of Electrical and Electronics Engineering
dc.departmentNational Magnetic Resonance Research Center (UMRAM)
dc.departmentGraduate Program in Neuroscience - Ph.D. / Sc.D.
dc.description.abstractDeep MRI reconstruction is commonly performed with conditional models that de-alias undersampled acquisitions to recover images consistent with fully-sampled data. Since conditional models are trained with knowledge of the imaging operator, they can show poor generalization across variable operators. Unconditional models instead learn generative image priors decoupled from the operator to improve reliability against domain shifts related to the imaging operator. Recent diffusion models are particularly promising given their high sample fidelity. Nevertheless, inference with a static image prior can perform suboptimally. Here we propose the first adaptive diffusion prior for MRI reconstruction, AdaDiff, to improve performance and reliability against domain shifts. AdaDiff leverages an efficient diffusion prior trained via adversarial mapping over large reverse diffusion steps. A two-phase reconstruction is executed following training: a rapid-diffusion phase that produces an initial reconstruction with the trained prior, and an adaptation phase that further refines the result by updating the prior to minimize data-consistency loss. Demonstrations on multi-contrast brain MRI clearly indicate that AdaDiff outperforms competing conditional and unconditional methods under domain shifts, and achieves superior or on par within-domain performance. © 2023 Elsevier B.V.
dc.identifier.doi10.1016/j.media.2023.102872
dc.identifier.issn1361-8415
dc.identifier.urihttps://hdl.handle.net/11693/114837
dc.language.isoen_US
dc.publisherElsevier B.V.
dc.relation.isversionofhttps://dx.doi.org/10.1016/j.media.2023.102872
dc.source.titleMedical Image Analysis
dc.subjectDiffusion
dc.subjectAdaptive
dc.subjectMRI
dc.subjectReconstruction
dc.subjectGenerative
dc.subjectImage prior
dc.titleAdaptive diffusion priors for accelerated MRI reconstruction
dc.typeArticle

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Adaptive_diffusion_priors_for_accelerated_MRI_reconstruction.pdf
Size:
3.25 MB
Format:
Adobe Portable Document Format
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
2.01 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Department of Electrical and Electronics Engineering
National Magnetic Resonance Research Center (UMRAM)