SD-YOLO: A Lightweight and High-Performance Deep Model for Small and Dense Object Detection

SD-YOLO: A Lightweight and High-Performance Deep Model for Small and Dense 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 SD-YOLO: A Lightweight and High-Performance Deep Model for Small and Dense Object Detection Phuc-Thinh Huynh, Minh-Thanh Le, Tran Duc Tan, Thien Huynh-The This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7255152/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 10 Nov, 2025 Read the published version in Signal, Image and Video Processing → Version 1 posted 5 You are reading this latest preprint version Abstract Object detection in remote sensing imagery from unmanned aerial vehicles (UAVs) is crucial yet challenging, demanding efficient algorithms for high accuracy and real-time performance despite complexities like small, dense, and occluded objects in intricate backgrounds. To address these challenges, we introduce SD-YOLO, an enhanced object detection model based on You Only Look Once version 8 (YOLOv8). SD-YOLO incorporates several key innovations. First, SD-YOLO optimizes the model for resource-constrained platforms by removing redundant low-resolution feature maps and integrating a tiny detection head, accordingly improving small object detection while significantly reducing parameters by around \(65%\). Second, SD-YOLO enhances feature extraction with the C2f-DMSC block, an advanced combination of a Dense Multi-Scale Convolution (DMSC) block and a transformer module, to effectively capture local and global features for improved object representation. Third, the Multi-Scale Convolutional Block Attention Module (MSCBAM) refines feature processing by emphasizing critical regions and expanding the receptive field. To serve diverse demands of performance and efficiency, we offer two versions of SD-YOLO for either efficiency or accuracy via channel scaling. Evaluations on VisDrone-2019 show SD-YOLOn achieves a mean average precision (mAP 0.5 ) of 35.8% a 2.2% improvement over YOLOv8n, while SD-YOLOs reaches 43.7% mAP 0.5 on VisDrone-2019 and 79.2% mAP 0.5 on LEVIR-Ship with 3.62 M parameters, thus demonstrating its effectiveness for small, dense object detection in remote sensing. object detection remote sensing deep learning aerial imagery unmanned aerial vehicles You Only Look Once. Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 10 Nov, 2025 Read the published version in Signal, Image and Video Processing → Version 1 posted Editorial decision: Revision requested 09 Aug, 2025 Reviewers invited by journal 09 Aug, 2025 Editor assigned by journal 31 Jul, 2025 Submission checks completed at journal 31 Jul, 2025 First submitted to journal 30 Jul, 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-7255152","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":498187976,"identity":"8362617a-73b2-41e3-a66b-4f5c26b564d2","order_by":0,"name":"Phuc-Thinh Huynh","email":"","orcid":"","institution":"Ho Chi Minh City University of Technology and Education","correspondingAuthor":false,"prefix":"","firstName":"Phuc-Thinh","middleName":"","lastName":"Huynh","suffix":""},{"id":498187978,"identity":"ae45d1ff-8444-4456-8444-70324d4f49ec","order_by":1,"name":"Minh-Thanh Le","email":"","orcid":"","institution":"Ho Chi Minh City University of Technology and Education","correspondingAuthor":false,"prefix":"","firstName":"Minh-Thanh","middleName":"","lastName":"Le","suffix":""},{"id":498187980,"identity":"9a8f8f81-1464-4a46-8639-840bf9b69c88","order_by":2,"name":"Tran Duc Tan","email":"","orcid":"","institution":"Phenikaa University","correspondingAuthor":false,"prefix":"","firstName":"Tran","middleName":"Duc","lastName":"Tan","suffix":""},{"id":498187981,"identity":"093b9717-9f4b-4e3b-b253-ca1feeb9f80b","order_by":3,"name":"Thien Huynh-The","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA20lEQVRIiWNgGAWjYDACCcYGBiCSM4CLHCCggweixcDYAK6asBYgAdSSuIFoLfbSzY2PC3f8Sd/OfjpN+kMFgxzfjQQ26QJ8tsgcbDaeecYgd2dP7jaJA2cYjCVBWmbgdVhimzRvm0Huhhu82yQOtjEkbriRwGzMg19L+2+glnQDsJZ/DPXEaGljBmpJgGhpYAAyEhgf49VyI7EZ6DBjww1ncjdbnDkmYTjzzMNGvFrYZ6Q//MzbJidvcPzsxhsVNTbyfMeTDxzGpwUdQKKJBA2jYBSMglEwCrABAPoyTd8XX+oSAAAAAElFTkSuQmCC","orcid":"","institution":"Ho Chi Minh City University of Technology and Education","correspondingAuthor":true,"prefix":"","firstName":"Thien","middleName":"","lastName":"Huynh-The","suffix":""}],"badges":[],"createdAt":"2025-07-30 16:53:26","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7255152/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7255152/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11760-025-04932-9","type":"published","date":"2025-11-10T15:57:07+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":96104962,"identity":"d7116d25-f8ce-42dd-92c3-4f8ca2c1f4c8","added_by":"auto","created_at":"2025-11-17 16:04:53","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1054711,"visible":true,"origin":"","legend":"","description":"","filename":"PhucThinhJ01Springer.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7255152/v1_covered_d1ab3ac1-48fb-4f7c-8b55-5d746cacdd82.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"SD-YOLO: A Lightweight and High-Performance Deep Model for Small and Dense Object Detection","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"object detection, remote sensing, deep learning, aerial imagery, unmanned aerial vehicles, You Only Look Once.","lastPublishedDoi":"10.21203/rs.3.rs-7255152/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7255152/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eObject detection in remote sensing imagery from unmanned aerial vehicles (UAVs) is crucial yet challenging, demanding efficient algorithms for high accuracy and real-time performance despite complexities like small, dense, and occluded objects in intricate backgrounds. To address these challenges, we introduce SD-YOLO, an enhanced object detection model based on You Only Look Once version 8 (YOLOv8). SD-YOLO incorporates several key innovations. First, SD-YOLO optimizes the model for resource-constrained platforms by removing redundant low-resolution feature maps and integrating a tiny detection head, accordingly improving small object detection while significantly reducing parameters by around \\(65%\\). Second, SD-YOLO enhances feature extraction with the C2f-DMSC block, an advanced combination of a Dense Multi-Scale Convolution (DMSC) block and a transformer module, to effectively capture local and global features for improved object representation. Third, the Multi-Scale Convolutional Block Attention Module (MSCBAM) refines feature processing by emphasizing critical regions and expanding the receptive field. To serve diverse demands of performance and efficiency, we offer two versions of SD-YOLO for either efficiency or accuracy via channel scaling. Evaluations on VisDrone-2019 show SD-YOLOn achieves a mean average precision (mAP\u003csub\u003e0.5\u003c/sub\u003e) of \u003cstrong\u003e35.8%\u003c/strong\u003e a \u003cstrong\u003e2.2%\u003c/strong\u003e improvement over YOLOv8n, while SD-YOLOs reaches \u003cstrong\u003e43.7%\u003c/strong\u003e mAP\u003csub\u003e0.5\u003c/sub\u003e on VisDrone-2019 and \u003cstrong\u003e79.2% \u003c/strong\u003emAP\u003csub\u003e0.5 \u003c/sub\u003eon LEVIR-Ship with \u003cstrong\u003e3.62\u003c/strong\u003eM parameters, thus demonstrating its effectiveness for small, dense object detection in remote sensing.\u003c/p\u003e","manuscriptTitle":"SD-YOLO: A Lightweight and High-Performance Deep Model for Small and Dense Object Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-12 02:46:48","doi":"10.21203/rs.3.rs-7255152/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-08-10T02:00:01+00:00","index":"","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-10T01:59:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-31T12:12:55+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-31T12:12:39+00:00","index":"","fulltext":""},{"type":"submitted","content":"Signal, Image and Video Processing","date":"2025-07-30T16:43:36+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"58979b70-fae4-4317-a793-3e12c25a67da","owner":[],"postedDate":"August 12th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-11-17T15:59:36+00:00","versionOfRecord":{"articleIdentity":"rs-7255152","link":"https://doi.org/10.1007/s11760-025-04932-9","journal":{"identity":"signal-image-and-video-processing","isVorOnly":false,"title":"Signal, Image and Video Processing"},"publishedOn":"2025-11-10 15:57:07","publishedOnDateReadable":"November 10th, 2025"},"versionCreatedAt":"2025-08-12 02:46:48","video":"","vorDoi":"10.1007/s11760-025-04932-9","vorDoiUrl":"https://doi.org/10.1007/s11760-025-04932-9","workflowStages":[]},"version":"v1","identity":"rs-7255152","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7255152","identity":"rs-7255152","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}