DIPLI: Deep Image Prior Lucky Imaging for BlindAstronomical Image Restoration

DIPLI: Deep Image Prior Lucky Imaging for BlindAstronomical Image Restoration | 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 DIPLI: Deep Image Prior Lucky Imaging for BlindAstronomical Image Restoration Suraj Singh, Anastasiia Batsheva, Oleg Rogov, Ahmed Bouridane This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7751684/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 15 Apr, 2026 Read the published version in Scientific Reports → Version 1 posted 11 You are reading this latest preprint version Abstract Modern image restoration and super-resolution methods utilize deep learning due to its superior performance compared to traditional algorithms. However, deep learning typically requires large training datasets, which are rarely available in astrophotography. Deep Image Prior (DIP) bypasses this constraint by performing blind training on a single image. Although effective in some cases, DIP often suffers from overfitting, artifact generation, and instability. To overcome these issues and improve general performance, this work proposes DIPLI - a framework that shifts from single-frame to multi-frame training using the Back Projection technique, combined with optical flow estimation via the TVNet model, and replaces deterministic predictions with unbiased Monte Carlo estimation obtained through Langevin dynamics. A comprehensive evaluation compares the method against Lucky Imaging, a classical computer vision technique still widely used in astronomical image reconstruction, DIP, the transformer-based model RVRT, and the diffusion-based model DiffIR2VR-Zero. Experiments on synthetic datasets demonstrate consistent improvements, with the method outperforming baselines for SSIM, PSNR, LPIPS, and DISTS metrics in the majority of cases. In addition to superior reconstruction quality, the model also requires far fewer input images than Lucky Imaging and is less prone to overfitting or artifact generation. Evaluation on real-world astronomical data, where domain shifts typically hinder generalization, shows that the method maintains high reconstruction quality, confirming practical robustness. Physical sciences/Engineering Physical sciences/Mathematics and computing Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 15 Apr, 2026 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 16 Mar, 2026 Reviews received at journal 14 Mar, 2026 Reviewers agreed at journal 08 Mar, 2026 Reviews received at journal 08 Mar, 2026 Reviewers agreed at journal 07 Mar, 2026 Reviewers agreed at journal 06 Mar, 2026 Reviewers invited by journal 14 Nov, 2025 Editor assigned by journal 07 Nov, 2025 Editor invited by journal 03 Nov, 2025 Submission checks completed at journal 01 Nov, 2025 First submitted to journal 01 Nov, 2025 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-7751684","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":549827444,"identity":"e390f51e-4532-45f8-8bb2-08cb85792bcf","order_by":0,"name":"Suraj Singh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/0lEQVRIie2RsWrDMBCGzwic5YesBw3NK6gEnIQG8iqelKUZurVLKHTIkgfIS+QNMggEngRZDV7SxVMI7uYxUkNLh9ruWKi+4YTEfdz9iCgQ+IMMX1xhWoEI0ZE0Ddw9blWk/lCMV4R0CtCpXA8vIubfKbx8qyakB+P+Lnuu7Qzz3mt5jPamOct2MWKfZbo9qwK5ApCNJZXNCuWKnCIgc5sUVBmAH/yGrYqoP5XH2ivDU6cSX6ccNgkh91PQrkhbxhOWxk3B6AbWZbEq4VQvmuOvlSj4aTWXB3v3Xmez297alFzp++bFHIL973D67Sn9ufOLqPK1rzvaAoFA4N9yAXMxTuaJpVIuAAAAAElFTkSuQmCC","orcid":"","institution":"Skolkovo Institute of Science and Technology","correspondingAuthor":true,"prefix":"","firstName":"Suraj","middleName":"","lastName":"Singh","suffix":""},{"id":549827445,"identity":"b5bac95e-0116-4a78-9f27-c0c541288c40","order_by":1,"name":"Anastasiia Batsheva","email":"","orcid":"","institution":"AIRI","correspondingAuthor":false,"prefix":"","firstName":"Anastasiia","middleName":"","lastName":"Batsheva","suffix":""},{"id":549827446,"identity":"f4abcf6a-a388-4a88-b33f-f4bded2334d3","order_by":2,"name":"Oleg Rogov","email":"","orcid":"","institution":"AIRI","correspondingAuthor":false,"prefix":"","firstName":"Oleg","middleName":"","lastName":"Rogov","suffix":""},{"id":549827447,"identity":"e75ce035-306d-431f-b8b5-969d82095059","order_by":3,"name":"Ahmed Bouridane","email":"","orcid":"","institution":"University of Sharjah","correspondingAuthor":false,"prefix":"","firstName":"Ahmed","middleName":"","lastName":"Bouridane","suffix":""}],"badges":[],"createdAt":"2025-09-30 12:53:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7751684/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7751684/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-026-47300-4","type":"published","date":"2026-04-15T15:57:01+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":96686499,"identity":"1ec5d329-9023-4785-8156-cbba6e305e06","added_by":"auto","created_at":"2025-11-25 05:32:37","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6105,"visible":true,"origin":"","legend":"","description":"","filename":"51611fb1713041a7b372c798181ea0b1.json","url":"https://assets-eu.researchsquare.com/files/rs-7751684/v1/2af49e7d6b1ee8664c756172.json"},{"id":107350700,"identity":"41b9e758-d8aa-4391-a670-13c006c80397","added_by":"auto","created_at":"2026-04-20 16:00:29","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1827335,"visible":true,"origin":"","legend":"","description":"","filename":"DIPLIupdated.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7751684/v1_covered_72f8a972-f819-46e2-b746-7a37ba53e0cb.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"DIPLI: Deep Image Prior Lucky Imaging for BlindAstronomical Image Restoration","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"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-7751684/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7751684/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Modern image restoration and super-resolution methods utilize deep learning due to its superior performance compared to traditional algorithms. However, deep learning typically requires large training datasets, which are rarely available in astrophotography. Deep Image Prior (DIP) bypasses this constraint by performing blind training on a single image. Although effective in some cases, DIP often suffers from overfitting, artifact generation, and instability. To overcome these issues and improve general performance, this work proposes DIPLI - a framework that shifts from single-frame to multi-frame training using the Back Projection technique, combined with optical flow estimation via the TVNet model, and replaces deterministic predictions with unbiased Monte Carlo estimation obtained through Langevin dynamics. A comprehensive evaluation compares the method against Lucky Imaging, a classical computer vision technique still widely used in astronomical image reconstruction, DIP, the transformer-based model RVRT, and the diffusion-based model DiffIR2VR-Zero. Experiments on synthetic datasets demonstrate consistent improvements, with the method outperforming baselines for SSIM, PSNR, LPIPS, and DISTS metrics in the majority of cases. In addition to superior reconstruction quality, the model also requires far fewer input images than Lucky Imaging and is less prone to overfitting or artifact generation. Evaluation on real-world astronomical data, where domain shifts typically hinder generalization, shows that the method maintains high reconstruction quality, confirming practical robustness.","manuscriptTitle":"DIPLI: Deep Image Prior Lucky Imaging for BlindAstronomical Image Restoration","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-25 05:32:33","doi":"10.21203/rs.3.rs-7751684/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-16T05:14:49+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-14T11:58:03+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"117821445116658310520863898940458380912","date":"2026-03-09T00:30:33+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-08T09:55:09+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"297693206614874368717835479628380097884","date":"2026-03-07T07:07:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"132403398874036714799983549247344501786","date":"2026-03-07T00:48:27+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-14T10:22:53+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-07T09:07:48+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-11-03T14:19:38+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-01T19:50:08+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-11-01T19:48:35+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":"079e90d2-4a51-4dde-99cb-4c9006b5a645","owner":[],"postedDate":"November 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":58498342,"name":"Physical sciences/Engineering"},{"id":58498343,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2026-04-20T16:00:12+00:00","versionOfRecord":{"articleIdentity":"rs-7751684","link":"https://doi.org/10.1038/s41598-026-47300-4","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2026-04-15 15:57:01","publishedOnDateReadable":"April 15th, 2026"},"versionCreatedAt":"2025-11-25 05:32:33","video":"","vorDoi":"10.1038/s41598-026-47300-4","vorDoiUrl":"https://doi.org/10.1038/s41598-026-47300-4","workflowStages":[]},"version":"v1","identity":"rs-7751684","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7751684","identity":"rs-7751684","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}