Conditional Injective Flows for Bayesian Imaging

Simple item page
dc.citation.firstpage224
dc.citation.journalTitleIEEE Transactions on Computational Imaging
dc.citation.lastpage237
dc.citation.volumeNumber9
dc.contributor.authorKhorashadizadeh, AmirEhsan
dc.contributor.authorKothari, Konik
dc.contributor.authorSalsi, Leonardo
dc.contributor.authorHarandi, Ali Aghababaei
dc.contributor.authorde Hoop, Maarten
dc.contributor.authorDokmanić, Ivan
dc.date.accessioned2023-03-23T14:10:32Z
dc.date.available2023-03-23T14:10:32Z
dc.date.issued2023
dc.description.abstractMost deep learning models for computational imaging regress a single reconstructed image. In practice, however, ill-posedness, nonlinearity, model mismatch, and noise often conspire to make such point estimates misleading or insufficient. The Bayesian approach models images and (noisy) measurements as jointly distributed random vectors and aims to approximate the posterior distribution of unknowns. Recent variational inference methods based on conditional normalizing flows are a promising alternative to traditional MCMC methods, but they come with drawbacks: excessive memory and compute demands for moderate to high resolution images and underwhelming performance on hard nonlinear problems. In this work, we propose C-Trumpets—conditional injective flows specifically designed for imaging problems, which greatly diminish these challenges. Injectivity reduces memory footprint and training time while low-dimensional latent space together with architectural innovations like fixed-volume-change layers and skip-connection revnet layers, C-Trumpets outperform regular conditional flow models on a variety of imaging and image restoration tasks, including limited-view CT and nonlinear inverse scattering, with a lower compute and memory budget. C-Trumpets enable fast approximation of point estimates like MMSE or MAP as well as physically-meaningful uncertainty quantification.
dc.identifier.citationKhorashadizadeh, AmirEhsan, Kothari, Konik, Salsi, Leonardo, et al.. "Conditional Injective Flows for Bayesian Imaging." <i>IEEE Transactions on Computational Imaging,</i> 9, (2023) IEEE: 224-237. https://doi.org/10.1109/TCI.2023.3248949.
dc.identifier.digitalConditional_Injective_Flows_for_Bayesian_Imaging
dc.identifier.doihttps://doi.org/10.1109/TCI.2023.3248949
dc.identifier.urihttps://hdl.handle.net/1911/114523
dc.language.isoeng
dc.publisherIEEE
dc.rightsThis work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.titleConditional Injective Flows for Bayesian Imaging
dc.typeJournal article
dc.type.dcmiText
dc.type.publicationpublisher version
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Conditional_Injective_Flows_for_Bayesian_Imaging.pdf
Size:
4.61 MB
Format:
Adobe Portable Document Format
Download
Collections
Faculty Publications
CAAM Publications