MHS-DETR: A Railway Frog Defect Detection Model Integrating Multi-Scale Edge Perception and Adaptive Structural Optimization

MHS-DETR: A Railway Frog Defect Detection Model Integrating Multi-Scale Edge Perception and Adaptive Structural Optimization | 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 MHS-DETR: A Railway Frog Defect Detection Model Integrating Multi-Scale Edge Perception and Adaptive Structural Optimization Xinyue Wang, Chenghai Yu, Chaoyue Wei This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7262085/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 As a critical component of the railway track system, the structural integrity of railway frogs is vital to the safe operation of trains. However, detecting surface defects on frogs remains challenging due to their small size, blurred edges, and complex backgrounds, which often lead to low detection accuracy in existing algorithms. To address these issues, this paper proposes a novel object detection model named MHS-DETR, which integrates multi-scale edge perception with adaptive optimization strategies. Built upon the RT-DETR framework, MHS-DETR introduces three systematic improvements: First, a lightweight Multi-scale Edge Information Extraction Network (MEIENet) is developed to significantly enhance the perception of small defect contours by integrating edge enhancement and dual-domain selection mechanisms. Second, a Hierarchical Attention Fusion Block (HAFB) is designed to improve the interaction and fusion of heterogeneous features through a dynamically weighted combination of local and global attention mechanisms, thereby enhancing the representation of multi-scale targets. Third, an adaptive loss function, SlideVarifocalLoss (SVL), is proposed to alleviate the imbalance between positive and negative samples by dynamically adjusting the loss weights of samples with ambiguous boundaries. To validate the effectiveness of the proposed model, a multi-scenario frog defect dataset named ForkDefect-2089—containing 2089 annotated images—was constructed. Experimental results show that MHS-DETR achieves an mAP50 of 70.3%, outperforming the baseline model by 4.2%, while reducing the number of parameters and computational cost by 15.6% and 7.4%, respectively. Moreover, it achieves a 2.3% performance improvement on the public NEU-DET dataset. Overall, the proposed MHS-DETR demonstrates superior robustness and generalization capabilities compared to mainstream detectors, offering strong technical support and practical value for achieving high-reliability, intelligent railway maintenance. RT-DETR Railway frog detects HAFB MEIENet MHS-DETR Full Text Additional Declarations No competing interests reported. Supplementary Files SupplementarymaterialsincludetheMEIEmodeldiagramandForkDefect2089datasetstatistics.pdf 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-7262085","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":498203840,"identity":"d66a698c-3356-4fe8-8257-f12ad3d5b839","order_by":0,"name":"Xinyue Wang","email":"","orcid":"","institution":"Zhejiang Sci-Tech University","correspondingAuthor":false,"prefix":"","firstName":"Xinyue","middleName":"","lastName":"Wang","suffix":""},{"id":498203841,"identity":"467cbff3-803f-47a0-bead-c44fcf357be9","order_by":1,"name":"Chenghai Yu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAvUlEQVRIiWNgGAWjYBACNvaGhAMf/9kw8xOthY/nwMODM9jS2CUbiNUiJ+H4+DAH22F+gwNEO0yCOeEwAw+ztPHx5A0MPyq2EaFFui3hcIEEm7HZmWcFjD1nbhOhReZMwuEZBjzJZjdyDJgZ24jRIpH/4TBPgkT95hnEa0lIOMxzwIDZQIJoLTwHEg7ObEhglgD65SBRfpFvb0j+8LHhPzN/e/LGBz8qiNCCBBKIjxqEFlJ1jIJRMApGwQgBAIrYPr1VR5vgAAAAAElFTkSuQmCC","orcid":"","institution":"Zhejiang Sci-Tech University","correspondingAuthor":true,"prefix":"","firstName":"Chenghai","middleName":"","lastName":"Yu","suffix":""},{"id":498203842,"identity":"3ab501f1-c792-49d6-933e-b3ff99a39e77","order_by":2,"name":"Chaoyue Wei","email":"","orcid":"","institution":"Zhejiang Sci-Tech University","correspondingAuthor":false,"prefix":"","firstName":"Chaoyue","middleName":"","lastName":"Wei","suffix":""}],"badges":[],"createdAt":"2025-07-31 12:08:52","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7262085/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7262085/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90361979,"identity":"2f392e30-f4ce-4e3d-9918-e703d08a6003","added_by":"auto","created_at":"2025-09-02 01:38:34","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1381155,"visible":true,"origin":"","legend":"","description":"","filename":"WangMHSDETRSIVP2025.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7262085/v1_covered_bb44713f-33b1-4096-a7e0-ebf792c08b40.pdf"},{"id":88881421,"identity":"1f427b74-976b-4ddc-8407-0f5addf5616c","added_by":"auto","created_at":"2025-08-12 11:09:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":154057,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementarymaterialsincludetheMEIEmodeldiagramandForkDefect2089datasetstatistics.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7262085/v1/208227c070cf93441a4e26a0.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"MHS-DETR: A Railway Frog Defect Detection Model Integrating Multi-Scale Edge Perception and Adaptive Structural Optimization","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":"RT-DETR, Railway frog detects, HAFB, MEIENet, MHS-DETR","lastPublishedDoi":"10.21203/rs.3.rs-7262085/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7262085/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAs a critical component of the railway track system, the structural integrity of railway frogs is vital to the safe operation of trains. However, detecting surface defects on frogs remains challenging due to their small size, blurred edges, and complex backgrounds, which often lead to low detection accuracy in existing algorithms. To address these issues, this paper proposes a novel object detection model named MHS-DETR, which integrates multi-scale edge perception with adaptive optimization strategies. Built upon the RT-DETR framework, MHS-DETR introduces three systematic improvements: First, a lightweight Multi-scale Edge Information Extraction Network (MEIENet) is developed to significantly enhance the perception of small defect contours by integrating edge enhancement and dual-domain selection mechanisms. Second, a Hierarchical Attention Fusion Block (HAFB) is designed to improve the interaction and fusion of heterogeneous features through a dynamically weighted combination of local and global attention mechanisms, thereby enhancing the representation of multi-scale targets. Third, an adaptive loss function, SlideVarifocalLoss (SVL), is proposed to alleviate the imbalance between positive and negative samples by dynamically adjusting the loss weights of samples with ambiguous boundaries. To validate the effectiveness of the proposed model, a multi-scenario frog defect dataset named ForkDefect-2089\u0026mdash;containing 2089 annotated images\u0026mdash;was constructed. Experimental results show that MHS-DETR achieves an mAP50 of 70.3%, outperforming the baseline model by 4.2%, while reducing the number of parameters and computational cost by 15.6% and 7.4%, respectively. Moreover, it achieves a 2.3% performance improvement on the public NEU-DET dataset. Overall, the proposed MHS-DETR demonstrates superior robustness and generalization capabilities compared to mainstream detectors, offering strong technical support and practical value for achieving high-reliability, intelligent railway maintenance.\u003c/p\u003e","manuscriptTitle":"MHS-DETR: A Railway Frog Defect Detection Model Integrating Multi-Scale Edge Perception and Adaptive Structural Optimization","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-12 11:09:52","doi":"10.21203/rs.3.rs-7262085/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":"55cd5b87-220b-4c87-8d33-e29bfb4d636f","owner":[],"postedDate":"August 12th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-09-02T01:38:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-12 11:09:52","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7262085","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7262085","identity":"rs-7262085","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}