CMH‑Net:a structured and optimized network for real-time steel surface defect detection

CMH‑Net:a structured and optimized network for real-time steel surface defect 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 CMH‑Net:a structured and optimized network for real-time steel surface defect detection Jie Li, Liangfu Li, Lingmei Ai This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7197826/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 12 Oct, 2025 Read the published version in Journal of Real-Time Image Processing → Version 1 posted 11 You are reading this latest preprint version Abstract Steel surface defect detection plays a pivotal role in modern industry, ensuring product quality and safety. However, significant challenges still remain, such as unstructured nature and multi-scale characteristics of defects. One of the key concerns is the balance between detection accuracy and real-time performance. To address the above challenges, this paper introduces CMH-Net, a lightweight one-stage detection model based on YOLOv12. CMH-Net is designed with several efficient modules that significantly enhance detection ability. The backbone incorporates a two-stage cascaded Haar wavelet downsampling module to improve unstructured feature extraction. Furthermore, a multi-scale selective Mamba-Like Linear Attention(MLLA) module is proposed for multi-scale feature fusion, improving detection accuracy for defects of varying scales. CMH-Net comprises and optimizes two structural modules with few parameters, ultimately enhancing feature representation, improving overall model performance and efficiency. Following this, a hybrid loss function is applied to enhance bounding box accuracy and model robustness. Experimental results on two public datasets for steel surface defect detection demonstrate that the prposed CMH-Net outperforms state-of-the-art methods in terms of the mAP50 metric (NEU-DET: 0.814, GC10-DET: 0.715). With only 3.6M parameters and 6.4ms inference time, CMH-Net achieves an optimal trade-off between detection accuracy and real-time efficiency. Surface defect detection Cascade extraction network Multi-scale feature fusion Real-time Deep learning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 12 Oct, 2025 Read the published version in Journal of Real-Time Image Processing → Version 1 posted Editorial decision: Revision requested 29 Aug, 2025 Reviews received at journal 28 Aug, 2025 Reviews received at journal 27 Aug, 2025 Reviews received at journal 26 Aug, 2025 Reviewers agreed at journal 07 Aug, 2025 Reviewers agreed at journal 07 Aug, 2025 Reviewers agreed at journal 07 Aug, 2025 Reviewers invited by journal 30 Jul, 2025 Editor assigned by journal 28 Jul, 2025 Submission checks completed at journal 28 Jul, 2025 First submitted to journal 23 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-7197826","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":494021986,"identity":"6ecb88bf-3ca6-47ce-90d5-fb8d77bc912e","order_by":0,"name":"Jie Li","email":"","orcid":"","institution":"Shaanxi Normal University","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Li","suffix":""},{"id":494021989,"identity":"85b57449-f458-46ba-b239-dc5302df17d9","order_by":1,"name":"Liangfu Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYBACPhBRwcDA2A8VYGwgpIUNRJwBqpzZQLKWDQeI1sLee/jFgYp7spvPnz32mIfBRnbDAeZnD/Bq4TmXZnHgTLHxtht56cY8DGnGGw6wmRvg1SKRY2b8sS0hcdsNHjNpHobDiRsO8LBJENJicBCoZXP/GZCW/0RpMX4A0rKBIQek5QARWnjOmDEcOJNgPONGjpnkHINk45mH2czwauFn7zH+cKAiQbYf6DCJNxV2sn3Hm5/h1QJ2G4INCipmAupBSj4QVjMKRsEoGAUjGgAAziRHsPU75M4AAAAASUVORK5CYII=","orcid":"","institution":"Shaanxi Normal University","correspondingAuthor":true,"prefix":"","firstName":"Liangfu","middleName":"","lastName":"Li","suffix":""},{"id":494021991,"identity":"9ae34465-bb77-495d-b77e-8d36f191fe6f","order_by":2,"name":"Lingmei Ai","email":"","orcid":"","institution":"Shaanxi Normal University","correspondingAuthor":false,"prefix":"","firstName":"Lingmei","middleName":"","lastName":"Ai","suffix":""}],"badges":[],"createdAt":"2025-07-23 15:08:25","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7197826/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7197826/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11554-025-01786-9","type":"published","date":"2025-10-12T15:58:09+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":93419776,"identity":"440d2b5d-7505-4083-903e-b96868c10c2c","added_by":"auto","created_at":"2025-10-13 16:07:26","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1206175,"visible":true,"origin":"","legend":"","description":"","filename":"CMHNetastructuredandoptimizednetworkforrealtimesteelsurfacedefectdetection.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7197826/v1_covered_448f2840-a8cd-4600-877e-6fdf6930659b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"CMH‑Net:a structured and optimized network for real-time steel surface defect detection","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"journal-of-real-time-image-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"rtip","sideBox":"Learn more about [Journal of Real-Time Image Processing](http://link.springer.com/journal/11554)","snPcode":"11554","submissionUrl":"https://submission.nature.com/new-submission/11554/3","title":"Journal of Real-Time Image Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Surface defect detection, Cascade extraction network, Multi-scale feature fusion, Real-time, Deep learning","lastPublishedDoi":"10.21203/rs.3.rs-7197826/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7197826/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Steel surface defect detection plays a pivotal role in modern industry, ensuring product quality and safety. However, significant challenges still remain, such as unstructured nature and multi-scale characteristics of defects. One of the key concerns is the balance between detection accuracy and real-time performance. To address the above challenges, this paper introduces CMH-Net, a lightweight one-stage detection model based on YOLOv12. CMH-Net is designed with several efficient modules that significantly enhance detection ability. The backbone incorporates a two-stage cascaded Haar wavelet downsampling module to improve unstructured feature extraction. Furthermore, a multi-scale selective Mamba-Like Linear Attention(MLLA) module is proposed for multi-scale feature fusion, improving detection accuracy for defects of varying scales. CMH-Net comprises and optimizes two structural modules with few parameters, ultimately enhancing feature representation, improving overall model performance and efficiency. Following this, a hybrid loss function is applied to enhance bounding box accuracy and model robustness. Experimental results on two public datasets for steel surface defect detection demonstrate that the prposed CMH-Net outperforms state-of-the-art methods in terms of the mAP50 metric (NEU-DET: 0.814, GC10-DET: 0.715). With only 3.6M parameters and 6.4ms inference time, CMH-Net achieves an optimal trade-off between detection accuracy and real-time efficiency.","manuscriptTitle":"CMH‑Net:a structured and optimized network for real-time steel surface defect detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-04 08:09:43","doi":"10.21203/rs.3.rs-7197826/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-08-29T13:28:18+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-28T20:42:20+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-27T04:12:43+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-26T21:09:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"179316747003581510923623609171383073415","date":"2025-08-07T21:33:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"247273945539219382158548316195048338902","date":"2025-08-07T17:09:43+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"301680209722323701744772792650053220833","date":"2025-08-07T10:27:28+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-07-30T05:28:31+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-28T09:41:12+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-28T09:40:06+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Real-Time Image Processing","date":"2025-07-23T15:00:41+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-real-time-image-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"rtip","sideBox":"Learn more about [Journal of Real-Time Image Processing](http://link.springer.com/journal/11554)","snPcode":"11554","submissionUrl":"https://submission.nature.com/new-submission/11554/3","title":"Journal of Real-Time Image Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"bd0cb6c2-c33f-405f-9ab0-9c0c8ed35d33","owner":[],"postedDate":"August 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-10-13T16:02:37+00:00","versionOfRecord":{"articleIdentity":"rs-7197826","link":"https://doi.org/10.1007/s11554-025-01786-9","journal":{"identity":"journal-of-real-time-image-processing","isVorOnly":false,"title":"Journal of Real-Time Image Processing"},"publishedOn":"2025-10-12 15:58:09","publishedOnDateReadable":"October 12th, 2025"},"versionCreatedAt":"2025-08-04 08:09:43","video":"","vorDoi":"10.1007/s11554-025-01786-9","vorDoiUrl":"https://doi.org/10.1007/s11554-025-01786-9","workflowStages":[]},"version":"v1","identity":"rs-7197826","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7197826","identity":"rs-7197826","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}