An ensemble-based deep learning method through multi-scale cross-attention training for cephalometric landmark localization on lateral X-ray images

An ensemble-based deep learning method through multi-scale cross-attention training for cephalometric landmark localization on lateral X-ray images | 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 Article An ensemble-based deep learning method through multi-scale cross-attention training for cephalometric landmark localization on lateral X-ray images Zhu Zhu, Xiaoling Gu, Liuling Dong, Yu Liu, Yunyi Wang, Shuhui Huang, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6105085/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 Cephalometric landmark localization constitutes a core step in the diagnosis and treatment planning of orthodontics, or-thognathic surgery, and maxillofacial surgery. However, the manual marking process is time-consuming, laborious, and subject to inter-observer differences. Despite the advancements made by deep learning-based methods (such as heatmap regression), existing single-scale feature extraction and two-stage strategies still encounter problems like limited accuracy, error accumulation, and low computational efficiency. To this end, this paper proposes a novel multi-scale cross-attention training framework and ensemble learning strategy, aiming to enhance the robustness and accuracy of landmark localization by collaboratively modeling local details and global context relationships. The multi-scale cross-attention mechanism captures the local details and global spatial dependencies of anatomical landmarks through cross-scale feature interaction, generating complementary and enhanced feature representations. The strategy of heterogeneous model ensemble learning can combine the cross-attention features of multiple network architectures to mitigate the performance fluctuations of a single model and optimize the balance of landmark prediction. On the ISBI dataset (19 landmarks) and the MICCAI 2023 CL-Detection Challenge dataset (38 landmarks), this method achieved leading performances of 82.58% SDR (1.44 mm MRE) and 77.95% SDR (1.73 mm MRE), respectively, surpassing the existing optimal methods. The ablation experiments further indicated that the multi-scale interaction and ensemble strategies both contributed improvements in SDR. This approach realizes high-precision fully automatic localization in complex multi-landmark scenarios, providing efficient and consistent quantitative analysis tools for clinical practice. Simultaneously, it offers a universal technical framework for multi-scale modeling and model ensemble in key point detection tasks of medical images. Health sciences/Anatomy/Oral anatomy Physical sciences/Engineering/Biomedical engineering Physical sciences/Mathematics and computing/Computer science 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-6105085","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":427051596,"identity":"f5b21c7a-ee93-434b-a115-35bf3032bd48","order_by":0,"name":"Zhu Zhu","email":"","orcid":"","institution":"National Clinical Research Center for Child Health, National Children's Regional Medical Center, Children's Hospital, Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Zhu","middleName":"","lastName":"Zhu","suffix":""},{"id":427051597,"identity":"246b5a35-168d-42c7-b4e0-6b6cc9e3c21c","order_by":1,"name":"Xiaoling Gu","email":"","orcid":"","institution":"Key Laboratory of Complex Systems Modeling and Simulation, Hangzhou Dianzi University School of Computer Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Xiaoling","middleName":"","lastName":"Gu","suffix":""},{"id":427051598,"identity":"efe9bca7-7b13-403a-aac7-31e974e1a87d","order_by":2,"name":"Liuling Dong","email":"","orcid":"","institution":"Key Laboratory of Complex Systems Modeling and Simulation, Hangzhou Dianzi University School of Computer Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Liuling","middleName":"","lastName":"Dong","suffix":""},{"id":427051599,"identity":"c9cf360a-b680-45b9-bd46-20282689cc30","order_by":3,"name":"Yu Liu","email":"","orcid":"","institution":"Key Laboratory of Complex Systems Modeling and Simulation, Hangzhou Dianzi University School of Computer Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Liu","suffix":""},{"id":427051601,"identity":"4a4aeace-bf7d-49cb-8a36-9db85f79c1eb","order_by":4,"name":"Yunyi Wang","email":"","orcid":"","institution":"National Clinical Research Center for Child Health, National Children's Regional Medical Center, Children's Hospital, Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Yunyi","middleName":"","lastName":"Wang","suffix":""},{"id":427051604,"identity":"46ab0a44-de88-4ecb-a7eb-636d7216d295","order_by":5,"name":"Shuhui Huang","email":"","orcid":"","institution":"National Clinical Research Center for Child Health, National Children's Regional Medical Center, Children's Hospital, Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Shuhui","middleName":"","lastName":"Huang","suffix":""},{"id":427051606,"identity":"52df03aa-a651-4ef6-a35a-c48727705569","order_by":6,"name":"Yuyue Xu","email":"","orcid":"","institution":"National Clinical Research Center for Child Health, National Children's Regional Medical Center, Children's Hospital, Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Yuyue","middleName":"","lastName":"Xu","suffix":""},{"id":427051608,"identity":"42cee3ae-4b3a-461e-a39f-c80ee5ad78d6","order_by":7,"name":"Bin Xu","email":"","orcid":"","institution":"National Clinical Research Center for Child Health, National Children's Regional Medical Center, Children's Hospital, Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Bin","middleName":"","lastName":"Xu","suffix":""},{"id":427051609,"identity":"802fed0b-5c4d-481a-ba5c-cb1f5741bd28","order_by":8,"name":"Gang Yu","email":"","orcid":"","institution":"National Clinical Research Center for Child Health, National Children's Regional Medical Center, Children's Hospital, Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Gang","middleName":"","lastName":"Yu","suffix":""},{"id":427051610,"identity":"888fb6c9-869e-4488-bbcf-bba5348b435d","order_by":9,"name":"Feng Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAt0lEQVRIiWNgGAWjYDAC5gOMBz4wsIGYBkRqYUtgODiDZC2HeSBMIrXIt/EYHLb5w5fYwN68TYKh5g5hLYwgLTk8bIkNPMfKJBiOPSOshVm+d8PhHAmgFokcMwnGhsOEtbCx8W44bGEA1CL/hkgtPCAtDAkgW3iI1CLBxv/hYM8BNuM2nrRii4RjRGiRb2NLfPDjzzHZfvbDG298qCFCCxQcg0RmAtEaGBhqSFA7CkbBKBgFIw4AAPxGNIbl8pLZAAAAAElFTkSuQmCC","orcid":"","institution":"National Clinical Research Center for Child Health, National Children's Regional Medical Center, Children's Hospital, Zhejiang University School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Feng","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-02-25 12:23:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6105085/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6105085/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95802820,"identity":"6a8f1cca-0512-4e48-93cb-3c27ce0537aa","added_by":"auto","created_at":"2025-11-13 08:28:40","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":559237,"visible":true,"origin":"","legend":"","description":"","filename":"scientificreportdigitalhealthcephalometriclandmarkdetection.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6105085/v1_covered_3c2b024a-a677-4de8-8cbb-1d03f9553f5a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"An ensemble-based deep learning method through multi-scale cross-attention training for cephalometric landmark localization on lateral X-ray images","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":"","lastPublishedDoi":"10.21203/rs.3.rs-6105085/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6105085/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Cephalometric landmark localization constitutes a core step in the diagnosis and treatment planning of orthodontics, or-thognathic surgery, and maxillofacial surgery. However, the manual marking process is time-consuming, laborious, and subject to inter-observer differences. Despite the advancements made by deep learning-based methods (such as heatmap regression), existing single-scale feature extraction and two-stage strategies still encounter problems like limited accuracy, error accumulation, and low computational efficiency. To this end, this paper proposes a novel multi-scale cross-attention training framework and ensemble learning strategy, aiming to enhance the robustness and accuracy of landmark localization by collaboratively modeling local details and global context relationships. The multi-scale cross-attention mechanism captures the local details and global spatial dependencies of anatomical landmarks through cross-scale feature interaction, generating complementary and enhanced feature representations. The strategy of heterogeneous model ensemble learning can combine the cross-attention features of multiple network architectures to mitigate the performance fluctuations of a single model and optimize the balance of landmark prediction. On the ISBI dataset (19 landmarks) and the MICCAI 2023 CL-Detection Challenge dataset (38 landmarks), this method achieved leading performances of 82.58% SDR (1.44 mm MRE) and 77.95% SDR (1.73 mm MRE), respectively, surpassing the existing optimal methods. The ablation experiments further indicated that the multi-scale interaction and ensemble strategies both contributed improvements in SDR. This approach realizes high-precision fully automatic localization in complex multi-landmark scenarios, providing efficient and consistent quantitative analysis tools for clinical practice. Simultaneously, it offers a universal technical framework for multi-scale modeling and model ensemble in key point detection tasks of medical images.","manuscriptTitle":"An ensemble-based deep learning method through multi-scale cross-attention training for cephalometric landmark localization on lateral X-ray images","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-12 15:13:08","doi":"10.21203/rs.3.rs-6105085/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":"dac1dea6-746d-4455-88ea-3552ed1d7f55","owner":[],"postedDate":"March 12th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":45502741,"name":"Health sciences/Anatomy/Oral anatomy"},{"id":45502742,"name":"Physical sciences/Engineering/Biomedical engineering"},{"id":45502743,"name":"Physical sciences/Mathematics and computing/Computer science"}],"tags":[],"updatedAt":"2025-11-13T07:08:09+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-12 15:13:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6105085","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6105085","identity":"rs-6105085","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}