SGDet-Light: Synergistic Global-Local Learning for Efficient Small Object Detection

SGDet-Light: Synergistic Global-Local Learning for Efficient Small Object Detection | 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 SGDet-Light: Synergistic Global-Local Learning for Efficient Small Object Detection Di WU, ZhongZheng Liu, ZiHan Chen, YuePing Xiao, XiaoLin Zhu, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7130158/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 To address the critical challenge of spatial information degradation in low-resolution small objects and prohibitive computational costs hindering mobile deployment, this work proposes SGDet-Light —a novel lightweight detection framework integrating Cross-layer Global Attention (CGA) and SandGlass Bottleneck Convolution (SGB). First, a hierarchical CGA mechanism is designed to establish inter-layer contextual dependencies, effectively suppressing feature redundancy while mitigating over-fitting across diverse scenarios. Second, an SandGlass bottleneck convolution (SGB) architecture with dual-path differentiable operators — Enhanced SandGlass Convolution (ESConv) for spatial-identity preservation in deep layers and Fused SandGlass Convolution (FSConv) for parameter efficiency in shallow layers to resolves gradient confusion and accelerates training convergence. Finally, we introduce an adaptive spatial feature fusion technique as the model’s multi-head predictor, en Experimental results demonstrate that our method outperforms state-of-the-art approaches on the MS COCO 2017 Val, Pascal VOC 2012 Val, and DOTA datasets, achieving improvements of 1.3% in AP, 1.4% in AP50, and 2.2% in [email protected] , while reducing parameters by 5.4% compared to the lightweight MobileNetV3.The framework achieves an optimal accuracy-efficiency trade-off, enabling real-time inference on edge devices. Small Object Detection Cross-layer Global Attention Mechanism SandGlass Bottleneck Convolution Adaptive Spatial Feature Fusion 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-7130158","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":490078414,"identity":"587def40-93e5-4868-a909-cf74c585643f","order_by":0,"name":"Di WU","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+0lEQVRIiWNgGAWjYHACgwMfDP7JsTHzf3zMwCABFiGo5eCMggPG/OwNxsZEa2Hm+HAgcWbPATNpmAh+9TeSNx5mMLjDuOFGQlp14Q4LOQb25m0SDDV38GhJKzhcYPCM2eBGwrHbM89IGDPwHCuTYDj2DKcWsxs5BodnGDCzGdxIbLvN2yaR2CCRYybB2HAYvxYeA2YeoAvZioFa6hvk3xCl5bCEZM8xNmaglgQGCR78WuzPPCs4OMMgzYCfvYdZGqjFsI0nrdgi4RhuLZLtyZs/fPhjU9/GzMP4mbetTp6f/fDGGx9qcGvBBGwgIoEEDaNgFIyCUTAKMAEACChWARgecHwAAAAASUVORK5CYII=","orcid":"","institution":"Hunan Engineering University","correspondingAuthor":true,"prefix":"","firstName":"Di","middleName":"","lastName":"WU","suffix":""},{"id":490078415,"identity":"37d8ac6c-0418-4f05-b534-e31cebdabee6","order_by":1,"name":"ZhongZheng Liu","email":"","orcid":"","institution":"Hunan Engineering University","correspondingAuthor":false,"prefix":"","firstName":"ZhongZheng","middleName":"","lastName":"Liu","suffix":""},{"id":490078416,"identity":"96994aaf-b598-4f6c-91bb-0b1e3a2ee431","order_by":2,"name":"ZiHan Chen","email":"","orcid":"","institution":"Hunan Engineering University","correspondingAuthor":false,"prefix":"","firstName":"ZiHan","middleName":"","lastName":"Chen","suffix":""},{"id":490078417,"identity":"762fddce-f3c6-4960-a563-c14336eb2844","order_by":3,"name":"YuePing Xiao","email":"","orcid":"","institution":"Hunan Engineering University","correspondingAuthor":false,"prefix":"","firstName":"YuePing","middleName":"","lastName":"Xiao","suffix":""},{"id":490078418,"identity":"bc045826-c321-441e-831e-f443cd34f9e2","order_by":4,"name":"XiaoLin Zhu","email":"","orcid":"","institution":"Hunan Engineering University","correspondingAuthor":false,"prefix":"","firstName":"XiaoLin","middleName":"","lastName":"Zhu","suffix":""},{"id":490078419,"identity":"0a724a1c-6bf7-45a1-a390-61813e53302f","order_by":5,"name":"Qin WAN","email":"","orcid":"","institution":"Hunan Engineering University","correspondingAuthor":false,"prefix":"","firstName":"Qin","middleName":"","lastName":"WAN","suffix":""}],"badges":[],"createdAt":"2025-07-15 11:38:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7130158/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7130158/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89227777,"identity":"fdfb3a8b-d25e-4969-b418-b358cd5998d6","added_by":"auto","created_at":"2025-08-17 13:16:42","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1256014,"visible":true,"origin":"","legend":"","description":"","filename":"20050629.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7130158/v1_covered_5402c09f-276f-4ead-b5e6-27d235fea3ce.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"SGDet-Light: Synergistic Global-Local Learning for Efficient Small Object Detection","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":"Small Object Detection, Cross-layer Global Attention Mechanism, SandGlass Bottleneck Convolution, Adaptive Spatial Feature Fusion","lastPublishedDoi":"10.21203/rs.3.rs-7130158/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7130158/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eTo address the critical challenge of spatial information degradation in low-resolution small objects and prohibitive computational costs hindering mobile deployment, this work proposes SGDet-Light \u0026mdash;a novel lightweight detection framework integrating Cross-layer Global Attention (CGA) and SandGlass Bottleneck Convolution (SGB). First, a hierarchical CGA mechanism is designed to establish inter-layer contextual dependencies, effectively suppressing feature redundancy while mitigating over-fitting across diverse scenarios. Second, an SandGlass bottleneck convolution (SGB) architecture with dual-path differentiable operators \u0026mdash; Enhanced SandGlass Convolution (ESConv) for spatial-identity preservation in deep layers and Fused SandGlass Convolution (FSConv) for parameter efficiency in shallow layers to resolves gradient confusion and accelerates training convergence. Finally, we introduce an adaptive spatial feature fusion technique as the model\u0026rsquo;s multi-head predictor, en Experimental results demonstrate that our method outperforms state-of-the-art approaches on the MS COCO 2017 Val, Pascal VOC 2012 Val, and DOTA datasets, achieving improvements of 1.3% in AP, 1.4% in AP50, and 2.2% in [email protected], while reducing parameters by 5.4% compared to the lightweight MobileNetV3.The framework achieves an optimal accuracy-efficiency trade-off, enabling real-time inference on edge devices.\u003c/p\u003e","manuscriptTitle":"SGDet-Light: Synergistic Global-Local Learning for Efficient Small Object Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-28 09:02:29","doi":"10.21203/rs.3.rs-7130158/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":"e9d99c35-7d6b-477e-87b9-974d47ab352c","owner":[],"postedDate":"July 28th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-08-17T13:08:33+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-28 09:02:29","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7130158","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7130158","identity":"rs-7130158","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}