Sampling-Efficient Unconditional Pure-Deblurring Diffusion Models via Noise-Augmented Generation

Sampling-Efficient Unconditional Pure-Deblurring Diffusion Models via Noise-Augmented Generation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Sampling-Efficient Unconditional Pure-Deblurring Diffusion Models via Noise-Augmented Generation Byung-Woo Hong, Simon Korman This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8997181/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Diffusion models have proven to be powerful image priors, supporting unconditional generation and a range of conditional tasks such as image restoration and text-guided synthesis. They learn to reverse a progressive noising process, capturing regularities of the data distribution. Nevertheless, diffusion models retain drawbacks that include slow sampling, high computational cost, and nontrivial architectural complexity.Recent work has explored replacing noise-based corruption with more structured degradations, most commonly progressive blurring, which aligns with the multi-scale structure of images and has been shown to support high-quality generation. However, sampling requires introducing stochasticity into the otherwise deterministic blurring process, typically by adding Gaussian noise, which causes the reverse dynamics to remain, in effect, a denoising process.Our blurring-based approach applies Gaussian noise, solely as a data augmentation, symmetrically to both inputs and targets, allowing the network to learn a pure and simple deblurring operation, with the noise used only to regularize training and enable stochastic sampling. This separation between blur and noise yields a lightweight generative process that produces high-quality samples using only a small number of sampling steps, without requiring dedicated fast-sampling mechanisms. Empirical results confirm the effectiveness of the proposed approach and its practical viability for generative modeling. Physical sciences/Engineering Physical sciences/Mathematics and computing Full Text Additional Declarations No competing interests reported. Supplementary Files generativemodelviadeblurringSRsupps.pdf Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 29 Mar, 2026 Editor assigned by journal 29 Mar, 2026 Editor invited by journal 09 Mar, 2026 Submission checks completed at journal 05 Mar, 2026 First submitted to journal 05 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8997181","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":614746513,"identity":"e5a4ff33-1231-4c5a-a2f1-57ca48575326","order_by":0,"name":"Byung-Woo Hong","email":"","orcid":"","institution":"Chung-Ang University","correspondingAuthor":false,"prefix":"","firstName":"Byung-Woo","middleName":"","lastName":"Hong","suffix":""},{"id":614746514,"identity":"f2b78d95-fad4-4d66-a901-87e3d13383c3","order_by":1,"name":"Simon Korman","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9ElEQVRIie2RsWrDMBRFXzEoi0CrQ0r8C9d4Lv0VG0Gz2JCxQwdNnky1BvoT/QTDA0/N7qGDs2TK1KFkMKV2Y+hUuWMGHXgIgQ7vXkTk8Vwj8nKshxFEWF+u6XZeSSYlmRTMK5n5UWhSyKFEFTcf7dP7xtpadOctIlpwR51Dwb7UL3lzLHZtuogrIDbyAc5gUDIJcsHFa0gilOhvDOXuLpEdlS/eQNVi2QP3Rp1m6u+rJChKHt6kYiWBzIQzW/DW6KB45njXZuXqFtBleETtDFZpDvJPjpTlZnnqcWeVPhzOj45gvwzFR8bfqf8leDwej+dPvgFI20f8iKfI2QAAAABJRU5ErkJggg==","orcid":"","institution":"University of Haifa","correspondingAuthor":true,"prefix":"","firstName":"Simon","middleName":"","lastName":"Korman","suffix":""}],"badges":[],"createdAt":"2026-02-28 17:39:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8997181/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8997181/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106093790,"identity":"192e4196-7bb6-48a9-86bd-4d5a2610f5a0","added_by":"auto","created_at":"2026-04-03 11:39:12","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2058618,"visible":true,"origin":"","legend":"","description":"","filename":"qfcxmrpyhxxvdnjmrnbkvdfzhhgsrdjk.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8997181/v1_covered_a35296e6-3584-4fba-90ff-537e7560fd19.pdf"},{"id":105973037,"identity":"b5492905-7fd4-490b-a3d0-f217c15eee7e","added_by":"auto","created_at":"2026-04-02 04:13:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":39130464,"visible":true,"origin":"","legend":"","description":"","filename":"generativemodelviadeblurringSRsupps.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8997181/v1/0f3d6f15806ef38517d9674d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Sampling-Efficient Unconditional Pure-Deblurring Diffusion Models via Noise-Augmented Generation","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8997181/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8997181/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Diffusion models have proven to be powerful image priors, supporting unconditional generation and a range of conditional tasks such as image restoration and text-guided synthesis. They learn to reverse a progressive noising process, capturing regularities of the data distribution. Nevertheless, diffusion models retain drawbacks that include slow sampling, high computational cost, and nontrivial architectural complexity.Recent work has explored replacing noise-based corruption with more structured degradations, most commonly progressive blurring, which aligns with the multi-scale structure of images and has been shown to support high-quality generation. However, sampling requires introducing stochasticity into the otherwise deterministic blurring process, typically by adding Gaussian noise, which causes the reverse dynamics to remain, in effect, a denoising process.Our blurring-based approach applies Gaussian noise, solely as a data augmentation, symmetrically to both inputs and targets, allowing the network to learn a pure and simple deblurring operation, with the noise used only to regularize training and enable stochastic sampling. This separation between blur and noise yields a lightweight generative process that produces high-quality samples using only a small number of sampling steps, without requiring dedicated fast-sampling mechanisms. Empirical results confirm the effectiveness of the proposed approach and its practical viability for generative modeling.","manuscriptTitle":"Sampling-Efficient Unconditional Pure-Deblurring Diffusion Models via Noise-Augmented Generation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-02 04:13:33","doi":"10.21203/rs.3.rs-8997181/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-03-29T05:54:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-29T05:54:24+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-09T19:43:09+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-05T17:47:08+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-03-05T11:00:30+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f1715b15-e7cd-4b2b-8e8d-d36d7d74c7f4","owner":[],"postedDate":"April 2nd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":65414046,"name":"Physical sciences/Engineering"},{"id":65414047,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2026-04-02T04:13:34+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-02 04:13:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8997181","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8997181","identity":"rs-8997181","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}