Physics-Guided Dataset Homogeneity Enables Universal Deep Learning Generalization in Scattering Media Imaging

Physics-Guided Dataset Homogeneity Enables Universal Deep Learning Generalization in Scattering Media Imaging | 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 Physics-Guided Dataset Homogeneity Enables Universal Deep Learning Generalization in Scattering Media Imaging Honglin Liu, Xuyu Zhang, Haofan Huang, Dawei Zhang, Songlin Zhuang, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7097373/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Deep learning (DL) is widely used in computational imaging for problems lacking analytic solutions, yet its generalization across datasets remains a critical challenge. Conventional wisdom attributes this limitation to feature-prior mismatches, but through physics-guided analysis of imaging through scattering media, we reveal a more fundamental cause: DL networks learn approximation s of the system’s true physical mapping ( \(\:{T}^{-1}\) ), constrained by the spatial-intensity distribution of training data. We demonstrate that enforcing the spacetime homogeneity—ensuring every point in the region of interest is equally and sufficiently trained—bridges the gap between learned mappings ( \(\:M\) ) and \(\:{T}^{-1}\) . By optimizing training datasets ( e.g ., transforming MNIST digits into grayscale-augmented variants), we achieve unprecedented cross-dataset generalization: networks trained on simple digits successfully reconstruct complex face images . This physics-guided framework not only overcomes generalization barriers in scattering imaging but also establishes a universal principle for designing robust DL architectures. Our work repositions DL from data-driven approximation to physics-simulating computation, unlocking reliable deployment in real-world applications. Physical sciences/Optics and photonics/Optical techniques/Imaging and sensing Physical sciences/Physics/Optical physics/Transformation optics deep learning mapping generalization imaging through scattering media homogeneity of spacetime Full Text Additional Declarations There is no conflict of interest Supplementary Files SupplementaryInformation.docx Supplementary Information for “Physics-Guided Dataset Homogeneity Enables Universal Deep Learning Generalization in Scattering Media” Cite Share Download PDF Status: Posted Version 1 posted 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-7097373","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":503962830,"identity":"58ac214a-ff6e-4d54-a9fa-39fe0ae3b571","order_by":0,"name":"Honglin Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwElEQVRIiWNgGAWjYBACA3bmhgMMPDYJYF5CATFamBmBWmTSEhjYQFoMiNTCwGBzGKKFgRgt5syMjYd5cs7n8ct3J354YMAgzy92AL8Wy2bGhsM8Z24XS7bxbpYAOsxw5uwEAg47DNTC23M7ccMx3g0gLQkGt4nS8u8cSMvmH8Rr4eE5ANKyjXhbDs7hSU6c2Za7zSLBQIIIvxxvPvzhDY9dYj/z2c03f1TYyPNLE9CCDiRIUz4KRsEoGAWjADsAAK9QRNiBH+KWAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-2473-9056","institution":"Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences","correspondingAuthor":true,"prefix":"","firstName":"Honglin","middleName":"","lastName":"Liu","suffix":""},{"id":503962831,"identity":"37d0a244-11f0-435e-874b-aff622ba4fb7","order_by":1,"name":"Xuyu Zhang","email":"","orcid":"","institution":"Shanghai Institute of Optics and Fine Mechanics","correspondingAuthor":false,"prefix":"","firstName":"Xuyu","middleName":"","lastName":"Zhang","suffix":""},{"id":503962832,"identity":"72ada52c-f6bf-4095-a145-062fcfced327","order_by":2,"name":"Haofan Huang","email":"","orcid":"","institution":"Department of Biomedical Engineering, The Hong Kong Polytechnic University","correspondingAuthor":false,"prefix":"","firstName":"Haofan","middleName":"","lastName":"Huang","suffix":""},{"id":503962833,"identity":"72dbb7a4-aa82-4f95-a5f9-3639f7450956","order_by":3,"name":"Dawei Zhang","email":"","orcid":"https://orcid.org/0000-0002-0841-7826","institution":"University of Shanghai for Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Dawei","middleName":"","lastName":"Zhang","suffix":""},{"id":503962834,"identity":"3966a565-b911-49b2-8377-2ed17e648cbc","order_by":4,"name":"Songlin Zhuang","email":"","orcid":"","institution":"University of Shanghai for Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Songlin","middleName":"","lastName":"Zhuang","suffix":""},{"id":503962835,"identity":"4cc8ce61-b4b9-4bb6-8931-7dac454633c4","order_by":5,"name":"shensheng han","email":"","orcid":"","institution":"Shanghai Institute of Optics and Fine Mechanics,Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"shensheng","middleName":"","lastName":"han","suffix":""},{"id":503962836,"identity":"64cd9e16-e07d-4cf7-8eef-b55394948629","order_by":6,"name":"Puxiang Lai","email":"","orcid":"https://orcid.org/0000-0003-4811-2012","institution":"The Hong Kong Polytechnic University","correspondingAuthor":false,"prefix":"","firstName":"Puxiang","middleName":"","lastName":"Lai","suffix":""}],"badges":[],"createdAt":"2025-07-11 03:40:35","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7097373/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7097373/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89789224,"identity":"4f28cf34-08cf-4187-a76a-12383f07e3e4","added_by":"auto","created_at":"2025-08-25 05:13:07","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1033319,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7097373/v1_covered_04da08fe-a65a-4422-bd29-b0ba36eeb301.pdf"},{"id":89788487,"identity":"ddd56bef-8409-41bf-b756-4aad18459790","added_by":"auto","created_at":"2025-08-25 04:57:00","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1190318,"visible":true,"origin":"","legend":"Supplementary Information for \u0026#x201C;Physics-Guided Dataset Homogeneity Enables Universal Deep Learning Generalization in Scattering Media\u0026#x201D;","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-7097373/v1/b312d2cfcfe01f9013324779.docx"}],"financialInterests":"There is no conflict of interest","formattedTitle":"Physics-Guided Dataset Homogeneity Enables Universal Deep Learning Generalization in Scattering Media Imaging","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"deep learning, mapping, generalization, imaging through scattering media, homogeneity of spacetime","lastPublishedDoi":"10.21203/rs.3.rs-7097373/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7097373/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDeep learning (DL) is widely used in computational imaging for problems lacking analytic solutions, yet its generalization across datasets remains a critical challenge. Conventional wisdom attributes this limitation to feature-prior mismatches, but through physics-guided analysis of imaging through scattering media, we reveal a more fundamental cause: DL networks learn approximation\u003cem\u003es\u003c/em\u003e of the system\u0026rsquo;s true physical mapping (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{T}^{-1}\\)\u003c/span\u003e\u003c/span\u003e), constrained by the spatial-intensity distribution of training data. We demonstrate that enforcing the spacetime homogeneity\u0026mdash;ensuring every point in the region of interest is equally and sufficiently trained\u0026mdash;bridges the gap between learned mappings (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:M\\)\u003c/span\u003e\u003c/span\u003e) and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{T}^{-1}\\)\u003c/span\u003e\u003c/span\u003e. By optimizing training datasets (\u003cem\u003ee.g\u003c/em\u003e., transforming MNIST digits into grayscale-augmented variants), we achieve unprecedented cross-dataset generalization: networks trained on \u003cem\u003esimple digits\u003c/em\u003e successfully reconstruct \u003cem\u003ecomplex face images\u003c/em\u003e. This physics-guided framework not only overcomes generalization barriers in scattering imaging but also establishes a universal principle for designing robust DL architectures. Our work repositions DL from data-driven approximation to physics-simulating computation, unlocking reliable deployment in real-world applications.\u003c/p\u003e","manuscriptTitle":"Physics-Guided Dataset Homogeneity Enables Universal Deep Learning Generalization in Scattering Media Imaging","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-25 04:56:55","doi":"10.21203/rs.3.rs-7097373/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"51fa80c5-6e82-4c21-b9d3-69b62d6ec889","owner":[],"postedDate":"August 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":53541403,"name":"Physical sciences/Optics and photonics/Optical techniques/Imaging and sensing"},{"id":53541404,"name":"Physical sciences/Physics/Optical physics/Transformation optics"}],"tags":[],"updatedAt":"2025-12-23T02:04:33+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-25 04:56:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7097373","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7097373","identity":"rs-7097373","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}