Low Illumination Target Detection Based on Information Aggregation and Distribution Mechanism

Low Illumination Target Detection Based on Information Aggregation and Distribution Mechanism | 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 Low Illumination Target Detection Based on Information Aggregation and Distribution Mechanism Xin Wang, Jian Li, yongshan Wang, qianhui Hua, Yi Shi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4517704/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 In low illumination environments, there are numerous challenges such as small targets, dense targets, occluded targets, and imbalanced sample distribution. Directly applying general object detection methods often fails to achieve ideal results. To address these challenges, this paper proposes an efficient object detection network, YOLO_LLD, for precise detection of targets in low illumination scenes. The algorithm is based on the YOLOv5s framework and introduces a cross-layer feature fusion method based on an information aggregation and distribution mechanism to mitigate information loss during cross-layer feature interactions. Additionally, the integration of dynamic sparse attention BiFormer constructs an efficient pyramid network architecture, reducing computational redundancy caused by the self-attention mechanism and enhancing the model's precision in detecting small targets.Inspired by the Inception structure, this paper designs the Multi-path Gradient Aggregation (MGA) structure, primarily aimed at ensuring better detail feature extraction from the perspective of gradient optimization under complex network models. Furthermore, a linear interval mapping mechanism is introduced into the bounding box regression loss function, enabling the network model to better focus on hard samples and further improve detection accuracy. Experimental results on the ExDark dataset demonstrate that, compared to YOLOv5, the mean average precision (mAP) is improved by 4.97%, indicating that the proposed method effectively enhances the performance of object detection in low illumination scenes. Object detection YOLOv5s Gather-Distribute Multi-path Gradient Aggregation attention mechanism 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-4517704","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":314903620,"identity":"f23fe609-daca-4862-a759-627c1449aa4d","order_by":0,"name":"Xin Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+0lEQVRIie3QMWrDMBSAYQmBsoiEbk8I3CuoGEKhgVzliUK7mFLokqmNMSiLD9DSS/QIDh285ADKVGtJDpDFWUpj6JYip1sG/aN4H3o8QmKxM2w0yPMGZ5CMBvPfp6qHyPKz0M3qOpVlN4onEO3u7YW3M6MdnkhIZSxBDkjW243ctyQZOqS7x4Cg86VtUMADfc/GSiBJpUOmXgOE0XyhEeCJqYyrw2LmwyFnIkA4oxZQg7Gy3sgWyUsvEbwjCKYEMobDYqj7CAhaaKwgBZGlN+IOrt5WvlAhMv3aer//fk6mde3X7WRyOaxvl7sQOfqVdFf8B4jFYrHYX/0A8KdMHFrmy90AAAAASUVORK5CYII=","orcid":"","institution":"Shaanxi University of Science and Technology","correspondingAuthor":true,"prefix":"","firstName":"Xin","middleName":"","lastName":"Wang","suffix":""},{"id":314903621,"identity":"e84a5f3f-f582-4971-b2a8-2fa6404f41eb","order_by":1,"name":"Jian Li","email":"","orcid":"","institution":"Shaanxi University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Li","suffix":""},{"id":314903622,"identity":"6e69e6be-ed92-4dce-924b-0a38712d848a","order_by":2,"name":"yongshan Wang","email":"","orcid":"","institution":"Shaanxi University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"yongshan","middleName":"","lastName":"Wang","suffix":""},{"id":314903623,"identity":"e7ce601b-4e00-458c-9382-e387639eb2b5","order_by":3,"name":"qianhui Hua","email":"","orcid":"","institution":"Shaanxi University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"qianhui","middleName":"","lastName":"Hua","suffix":""},{"id":314903624,"identity":"5176825d-bb99-43b1-b63b-365c568c5a2d","order_by":4,"name":"Yi Shi","email":"","orcid":"","institution":"Shaanxi University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"","lastName":"Shi","suffix":""}],"badges":[],"createdAt":"2024-06-02 16:08:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4517704/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4517704/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82224865,"identity":"fdeddd58-4f2f-4b3b-af88-991c99d996ff","added_by":"auto","created_at":"2025-05-08 03:46:54","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1144662,"visible":true,"origin":"","legend":"","description":"","filename":"LowIlluminationTargetDetectionBasedonInformationAggregationandDistributionMechanism.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4517704/v1_covered_7a7bf3e3-dc0d-4ade-97da-07d1dde92128.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Low Illumination Target Detection Based on Information Aggregation and Distribution Mechanism","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":"Object detection, YOLOv5s, Gather-Distribute, Multi-path Gradient Aggregation, attention mechanism","lastPublishedDoi":"10.21203/rs.3.rs-4517704/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4517704/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn low illumination environments, there are numerous challenges such as small targets, dense targets, occluded targets, and imbalanced sample distribution. Directly applying general object detection methods often fails to achieve ideal results. To address these challenges, this paper proposes an efficient object detection network, YOLO_LLD, for precise detection of targets in low illumination scenes. The algorithm is based on the YOLOv5s framework and introduces a cross-layer feature fusion method based on an information aggregation and distribution mechanism to mitigate information loss during cross-layer feature interactions. Additionally, the integration of dynamic sparse attention BiFormer constructs an efficient pyramid network architecture, reducing computational redundancy caused by the self-attention mechanism and enhancing the model's precision in detecting small targets.Inspired by the Inception structure, this paper designs the Multi-path Gradient Aggregation (MGA) structure, primarily aimed at ensuring better detail feature extraction from the perspective of gradient optimization under complex network models. Furthermore, a linear interval mapping mechanism is introduced into the bounding box regression loss function, enabling the network model to better focus on hard samples and further improve detection accuracy. Experimental results on the ExDark dataset demonstrate that, compared to YOLOv5, the mean average precision (mAP) is improved by 4.97%, indicating that the proposed method effectively enhances the performance of object detection in low illumination scenes.\u003c/p\u003e","manuscriptTitle":"Low Illumination Target Detection Based on Information Aggregation and Distribution Mechanism","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-17 04:01:22","doi":"10.21203/rs.3.rs-4517704/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":"03974dab-cb60-4a74-938a-bc3439970293","owner":[],"postedDate":"June 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-05-08T03:38:44+00:00","versionOfRecord":[],"versionCreatedAt":"2024-06-17 04:01:22","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4517704","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4517704","identity":"rs-4517704","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}