Dual-branch Feature Fusion Network for image denoising

Dual-branch Feature Fusion Network for image denoising | 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 Research Article Dual-branch Feature Fusion Network for image denoising Lijun Gao, Xiao Jin, Youzhi Zhang, Suran Wang, Zeyang Sun, Jiehong Wu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5659599/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 The goal of image denoising is to reduce or eliminate noise in images, thereby restoring and enhancing the true details and quality of the images. This process is essential for improving the accuracy of image analysis, enhancing image recognition, and optimizing visual effects. Despite the widespread application and notable success of deep convolutional neural networks (CNNs) in image denoising tasks, existing methods often encounter challenges such as overfitting and limited flexibility. To address these issues, this paper proposes a Dual-Branch Denoising Network (DFFNet) designed to enhance feature extraction capabilities and improve adaptability to noise. Specifically, the proposed DFFNet comprises two distinct parallel branches. In the first branch, a novel Multi-Scale Convolutional Block (MCB) is introduced, which significantly increases the receptive field while simultaneously preserving local details and the overall structure of the image. In the second branch, a Flexible Pooling Attention mechanism (FCA) is designed to reduce computational complexity while enhancing the effectiveness of feature representation. The two branches are then weighted and fused, achieving the dual objectives of noise removal and the preservation of critical image details, thereby improving overall image quality. Extensive experiments on multiple datasets demonstrate that the proposed DFFNet achieves superior denoising performance. Image denoising Dual-Branch Convolutional Network Multi-Scale Convolutional Attention mechanis Full Text Additional Declarations No competing interests reported. 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-5659599","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":391472420,"identity":"0320566f-8116-4dc7-9c04-1f138e0b4476","order_by":0,"name":"Lijun Gao","email":"","orcid":"","institution":"Shenyang Aerospace University","correspondingAuthor":false,"prefix":"","firstName":"Lijun","middleName":"","lastName":"Gao","suffix":""},{"id":391472424,"identity":"18779745-4805-4e95-9855-30d897b5e203","order_by":1,"name":"Xiao Jin","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1ElEQVRIiWNgGAWjYBACxgYEs/FBQkUNSVqYDxs8OHOMJAvZ0iQftjATVsc8u/nYxx8Vh+XM23vMKhIb2Bj427sT8DtszrHk2TxnDhvLnDljdiNxhwyDxJmzG/BrmZFjzMzYlpY4QyIHqOUMG4OBRC4hLfmfGX/+S6ufIf/GrCCxjZkYLTnMDLwNNgkSEmxpDMRpmXPMmJnnmI3hDJ7kwxIJZ47xEPSL4ezmx4w/aiTkJdgPNgKDrkaOv72XgJYZaAI8eJWDgLwEQSWjYBSMglEw4gEAG0xHJ5M/ytwAAAAASUVORK5CYII=","orcid":"","institution":"Shenyang Aerospace University","correspondingAuthor":true,"prefix":"","firstName":"Xiao","middleName":"","lastName":"Jin","suffix":""},{"id":391472426,"identity":"82ca2aec-9205-42b3-94bd-4ff93376e9df","order_by":2,"name":"Youzhi Zhang","email":"","orcid":"","institution":"Shenyang Aerospace University","correspondingAuthor":false,"prefix":"","firstName":"Youzhi","middleName":"","lastName":"Zhang","suffix":""},{"id":391472427,"identity":"e320dd5c-7208-4114-81ee-256262bd4de1","order_by":3,"name":"Suran Wang","email":"","orcid":"","institution":"Shenyang Aerospace University","correspondingAuthor":false,"prefix":"","firstName":"Suran","middleName":"","lastName":"Wang","suffix":""},{"id":391472429,"identity":"bd0b9eec-7bbf-4e63-9df0-0df18f72c53c","order_by":4,"name":"Zeyang Sun","email":"","orcid":"","institution":"Shenyang Aerospace University","correspondingAuthor":false,"prefix":"","firstName":"Zeyang","middleName":"","lastName":"Sun","suffix":""},{"id":391472437,"identity":"f6a47451-a78f-4b92-b06b-fd585eb7c5bf","order_by":5,"name":"Jiehong Wu","email":"","orcid":"","institution":"Shenyang Aerospace University","correspondingAuthor":false,"prefix":"","firstName":"Jiehong","middleName":"","lastName":"Wu","suffix":""}],"badges":[],"createdAt":"2024-12-17 08:08:28","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5659599/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5659599/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":75197422,"identity":"2326927b-b6a0-420d-8100-531151594785","added_by":"auto","created_at":"2025-01-31 21:31:28","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2290491,"visible":true,"origin":"","legend":"","description":"","filename":"SpringerNatureLaTeXTemplate.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5659599/v1_covered_2994911a-b2d4-4bbb-9bdf-966d02a8702b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Dual-branch Feature Fusion Network for image denoising","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"Image denoising, Dual-Branch Convolutional Network, Multi-Scale Convolutional, Attention mechanis","lastPublishedDoi":"10.21203/rs.3.rs-5659599/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5659599/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e The goal of image denoising is to reduce or eliminate noise in images, thereby restoring and enhancing the true details and quality of the images. This process is essential for improving the accuracy of image analysis, enhancing image recognition, and optimizing visual effects. Despite the widespread application and notable success of deep convolutional neural networks (CNNs) in image denoising tasks, existing methods often encounter challenges such as overfitting and limited flexibility. To address these issues, this paper proposes a Dual-Branch Denoising Network (DFFNet) designed to enhance feature extraction capabilities and improve adaptability to noise. Specifically, the proposed DFFNet comprises two distinct parallel branches. In the first branch, a novel Multi-Scale Convolutional Block (MCB) is introduced, which significantly increases the receptive field while simultaneously preserving local details and the overall structure of the image. In the second branch, a Flexible Pooling Attention mechanism (FCA) is designed to reduce computational complexity while enhancing the effectiveness of feature representation. The two branches are then weighted and fused, achieving the dual objectives of noise removal and the preservation of critical image details, thereby improving overall image quality. Extensive experiments on multiple datasets demonstrate that the proposed DFFNet achieves superior denoising performance.\u003c/p\u003e","manuscriptTitle":"Dual-branch Feature Fusion Network for image denoising","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-19 17:34:41","doi":"10.21203/rs.3.rs-5659599/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":"ea173074-a3cc-474a-972d-d6cbd4ce8d76","owner":[],"postedDate":"December 19th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-01-31T21:23:20+00:00","versionOfRecord":[],"versionCreatedAt":"2024-12-19 17:34:41","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5659599","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5659599","identity":"rs-5659599","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}