Matches in SemOpenAlex for { <https://semopenalex.org/work/W4206734100> ?p ?o ?g. }
- W4206734100 endingPage "250" @default.
- W4206734100 startingPage "239" @default.
- W4206734100 abstract "Reliable MRI is crucial for accurate interpretation in therapeutic and diagnostic tasks. However, undersampling during MRI acquisition as well as the overparameterized and non-transparent nature of deep learning (DL) leaves substantial uncertainty about the accuracy of DL reconstruction. With this in mind, this study aims to quantify the uncertainty in image recovery with DL models. To this end, we first leverage variational autoencoders (VAEs) to develop a probabilistic reconstruction scheme that maps out (low-quality) short scans with aliasing artifacts to the diagnostic-quality ones. The VAE encodes the acquisition uncertainty in a latent code and naturally offers a posterior of the image from which one can generate pixel variance maps using Monte-Carlo sampling. Accurately predicting risk requires knowledge of the bias as well, for which we leverage Stein's Unbiased Risk Estimator (SURE) as a proxy for mean-squared-error (MSE). A range of empirical experiments is performed for Knee MRI reconstruction under different training losses (adversarial and pixel-wise) and unrolled recurrent network architectures. Our key observations indicate that: 1) adversarial losses introduce more uncertainty; and 2) recurrent unrolled nets reduce the prediction uncertainty and risk." @default.
- W4206734100 created "2022-01-26" @default.
- W4206734100 creator A5008348052 @default.
- W4206734100 creator A5046357395 @default.
- W4206734100 creator A5060755739 @default.
- W4206734100 creator A5070743328 @default.
- W4206734100 date "2021-01-01" @default.
- W4206734100 modified "2023-10-15" @default.
- W4206734100 title "Uncertainty Quantification in Deep MRI Reconstruction" @default.
- W4206734100 cites W1992285975 @default.
- W4206734100 cites W2007203285 @default.
- W4206734100 cites W2023981146 @default.
- W4206734100 cites W2040468466 @default.
- W4206734100 cites W2098550720 @default.
- W4206734100 cites W2113920645 @default.
- W4206734100 cites W2129925362 @default.
- W4206734100 cites W2145568341 @default.
- W4206734100 cites W2559597482 @default.
- W4206734100 cites W2611467245 @default.
- W4206734100 cites W2617302333 @default.
- W4206734100 cites W2631883531 @default.
- W4206734100 cites W2754887662 @default.
- W4206734100 cites W2778924750 @default.
- W4206734100 cites W2791621240 @default.
- W4206734100 cites W2808312419 @default.
- W4206734100 cites W2883105305 @default.
- W4206734100 cites W2962821226 @default.
- W4206734100 cites W2963135377 @default.
- W4206734100 cites W3092596753 @default.
- W4206734100 cites W3102722370 @default.
- W4206734100 cites W4249760698 @default.
- W4206734100 doi "https://doi.org/10.1109/tmi.2020.3025065" @default.
- W4206734100 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/32956045" @default.
- W4206734100 hasPublicationYear "2021" @default.
- W4206734100 type Work @default.
- W4206734100 citedByCount "45" @default.
- W4206734100 countsByYear W42067341002020 @default.
- W4206734100 countsByYear W42067341002021 @default.
- W4206734100 countsByYear W42067341002022 @default.
- W4206734100 countsByYear W42067341002023 @default.
- W4206734100 crossrefType "journal-article" @default.
- W4206734100 hasAuthorship W4206734100A5008348052 @default.
- W4206734100 hasAuthorship W4206734100A5046357395 @default.
- W4206734100 hasAuthorship W4206734100A5060755739 @default.
- W4206734100 hasAuthorship W4206734100A5070743328 @default.
- W4206734100 hasBestOaLocation W42067341002 @default.
- W4206734100 hasConcept C105795698 @default.
- W4206734100 hasConcept C108583219 @default.
- W4206734100 hasConcept C11413529 @default.
- W4206734100 hasConcept C115961682 @default.
- W4206734100 hasConcept C119857082 @default.
- W4206734100 hasConcept C123614077 @default.
- W4206734100 hasConcept C136536468 @default.
- W4206734100 hasConcept C139945424 @default.
- W4206734100 hasConcept C141379421 @default.
- W4206734100 hasConcept C153083717 @default.
- W4206734100 hasConcept C154945302 @default.
- W4206734100 hasConcept C185429906 @default.
- W4206734100 hasConcept C31972630 @default.
- W4206734100 hasConcept C32230216 @default.
- W4206734100 hasConcept C33923547 @default.
- W4206734100 hasConcept C41008148 @default.
- W4206734100 hasConcept C49937458 @default.
- W4206734100 hasConcept C55020928 @default.
- W4206734100 hasConceptScore W4206734100C105795698 @default.
- W4206734100 hasConceptScore W4206734100C108583219 @default.
- W4206734100 hasConceptScore W4206734100C11413529 @default.
- W4206734100 hasConceptScore W4206734100C115961682 @default.
- W4206734100 hasConceptScore W4206734100C119857082 @default.
- W4206734100 hasConceptScore W4206734100C123614077 @default.
- W4206734100 hasConceptScore W4206734100C136536468 @default.
- W4206734100 hasConceptScore W4206734100C139945424 @default.
- W4206734100 hasConceptScore W4206734100C141379421 @default.
- W4206734100 hasConceptScore W4206734100C153083717 @default.
- W4206734100 hasConceptScore W4206734100C154945302 @default.
- W4206734100 hasConceptScore W4206734100C185429906 @default.
- W4206734100 hasConceptScore W4206734100C31972630 @default.
- W4206734100 hasConceptScore W4206734100C32230216 @default.
- W4206734100 hasConceptScore W4206734100C33923547 @default.
- W4206734100 hasConceptScore W4206734100C41008148 @default.
- W4206734100 hasConceptScore W4206734100C49937458 @default.
- W4206734100 hasConceptScore W4206734100C55020928 @default.
- W4206734100 hasFunder F4320332161 @default.
- W4206734100 hasIssue "1" @default.
- W4206734100 hasLocation W42067341001 @default.
- W4206734100 hasLocation W42067341002 @default.
- W4206734100 hasLocation W42067341003 @default.
- W4206734100 hasLocation W42067341004 @default.
- W4206734100 hasOpenAccess W4206734100 @default.
- W4206734100 hasPrimaryLocation W42067341001 @default.
- W4206734100 hasRelatedWork W2795261237 @default.
- W4206734100 hasRelatedWork W3014300295 @default.
- W4206734100 hasRelatedWork W3164822677 @default.
- W4206734100 hasRelatedWork W4223943233 @default.
- W4206734100 hasRelatedWork W4225161397 @default.
- W4206734100 hasRelatedWork W4312200629 @default.
- W4206734100 hasRelatedWork W4360585206 @default.
- W4206734100 hasRelatedWork W4364306694 @default.