CSP-HD: Confidence-Sensitive Progressive Hierarchical Distillation in Cross-View Geo-Localization | 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 CSP-HD: Confidence-Sensitive Progressive Hierarchical Distillation in Cross-View Geo-Localization Hai Yang, Min Xu, Zhihong Xu, Chaoyu Zhu, Xiaochen Li, Jing Ye, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8543337/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract Cross-view geo-localization (CVGL) enhances drone positioning by correlating aerial images with GPS-labeled satellite imagery.However, as data progresses through the network layers, the extracted representations become more semantically meaningful while simultaneously losing fine-grained spatial information.Moreover, the vast discrepancy in data volume between drone images and satellite images leads to uneven convergence during the training phase.Motivated by the knowledge distillation technique, we propose Confidence-Sensitive Progressive Hierarchical Distillation (CSP-HD) to enhance the effectiveness and efficiency of CVGL.Hierarchical Consistency Distillation (HCD) leverages the power of inverse self-distillation to simultaneously extract semantic features and spatial details.Progressive Adaptive Loss Weighting (PALW) dynamically adjusts the influence of HCD to alleviate short-term fluctuation during the training phase. Confidence-Sensitive Data Alignment (CSDA) dynamically adjusts the learning process based on the level of confidence to mitigate the data imbalance.Experimental results on public datasets demonstrate that our approach improves effectiveness by a margin of 1.50% to 1.84% in terms of Recall@1 and 1.50% to 1.65% in terms of average precision (AP) compared to prevailing approaches. Project page: https://1781950192.github.io/CSP-HD/}{https://1781950192.github.io/CSP-HD/. Hard Sample Mining Data Balance Cross-View Geo-Localization Computer Vision Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 11 Feb, 2026 Reviewers agreed at journal 03 Feb, 2026 Reviewers invited by journal 03 Feb, 2026 Editor assigned by journal 27 Jan, 2026 Submission checks completed at journal 10 Jan, 2026 First submitted to journal 07 Jan, 2026 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-8543337","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":584804093,"identity":"6797ce64-d3d8-4137-ade4-0f66c21c60ca","order_by":0,"name":"Hai Yang","email":"","orcid":"","institution":"China Tower Corporation Limited Zhejiang Branch","correspondingAuthor":false,"prefix":"","firstName":"Hai","middleName":"","lastName":"Yang","suffix":""},{"id":584804097,"identity":"ce9d83d7-29dc-4c96-87cc-ecfac8785d93","order_by":1,"name":"Min Xu","email":"","orcid":"","institution":"China Tower Corporation Limited Zhejiang Branch","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"Xu","suffix":""},{"id":584804099,"identity":"56e66eab-c8ed-49b7-bae7-f87fdc78d25b","order_by":2,"name":"Zhihong Xu","email":"","orcid":"","institution":"China Tower Corporation Limited Zhejiang Branch","correspondingAuthor":false,"prefix":"","firstName":"Zhihong","middleName":"","lastName":"Xu","suffix":""},{"id":584804105,"identity":"6590dde9-7090-48f5-8c43-f20e665fae9b","order_by":3,"name":"Chaoyu Zhu","email":"","orcid":"","institution":"China Tower Corporation Limited Zhejiang Branch","correspondingAuthor":false,"prefix":"","firstName":"Chaoyu","middleName":"","lastName":"Zhu","suffix":""},{"id":584804107,"identity":"3ced6455-bbb0-445e-80c4-01a88ae9e67e","order_by":4,"name":"Xiaochen Li","email":"","orcid":"","institution":"China Tower Corporation Limited Zhejiang Branch","correspondingAuthor":false,"prefix":"","firstName":"Xiaochen","middleName":"","lastName":"Li","suffix":""},{"id":584804108,"identity":"128759b2-4a0e-4acf-ae48-1943b94be115","order_by":5,"name":"Jing Ye","email":"","orcid":"","institution":"China Tower Corporation Limited Zhejiang Branch","correspondingAuthor":false,"prefix":"","firstName":"Jing","middleName":"","lastName":"Ye","suffix":""},{"id":584804109,"identity":"52346872-66d1-4114-b227-fbe57ad433b7","order_by":6,"name":"Chen Yang","email":"","orcid":"","institution":"China Tower Corporation Limited Zhejiang Branch","correspondingAuthor":false,"prefix":"","firstName":"Chen","middleName":"","lastName":"Yang","suffix":""},{"id":584804110,"identity":"eaec002b-adad-4dd4-9536-a23567661ba9","order_by":7,"name":"Guolong Xu","email":"","orcid":"","institution":"Zhejiang Gongshang University","correspondingAuthor":false,"prefix":"","firstName":"Guolong","middleName":"","lastName":"Xu","suffix":""},{"id":584804112,"identity":"bb2456ff-5166-4eab-834d-485435b41433","order_by":8,"name":"Yan Tian","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA60lEQVRIiWNgGAWjYDACZgglx8BwAMxgbCCkgweqxZgELVA6EaaSsBZ7duZnj3lq7qTPbzx++TMPg43shgPMzx7gdxibuTHPsWe5jQ1nyqR5GNKMNxxgMzfAr4XBTJqH7XBuM8OZNGYehsOJGw7wsEng18L+TZrn3+F0NoYzyUCH/SdGC4+ZNG/b4QQehuMHgA47QISWwzxlknP7DhvOYDjDJjnHINl45mE2M7xa2PuPb5N48+2wvPyM448/vKmwk+073vwMrxYQYAJHjsQZYDiBgoqZkHogYPwBIvnbHxChdhSMglEwCkYiAAD3yET7Gv4J2QAAAABJRU5ErkJggg==","orcid":"","institution":"Zhejiang Gongshang University","correspondingAuthor":true,"prefix":"","firstName":"Yan","middleName":"","lastName":"Tian","suffix":""}],"badges":[],"createdAt":"2026-01-07 15:38:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8543337/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8543337/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102295322,"identity":"e646015c-0215-47eb-a26a-e688c03bab81","added_by":"auto","created_at":"2026-02-10 10:10:54","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2333050,"visible":true,"origin":"","legend":"","description":"","filename":"CSPHDConfidenceSensitiveProgressiveHierarchicalDistillationinCrossViewGeoLocalization.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8543337/v1_covered_fff7ab2d-bf7e-4ecf-a240-de4f0acfb619.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"CSP-HD: Confidence-Sensitive Progressive Hierarchical Distillation in Cross-View Geo-Localization","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"multimedia-systems","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"mmsj","sideBox":"Learn more about [Multimedia Systems](http://link.springer.com/journal/530)","snPcode":"530","submissionUrl":"https://submission.nature.com/new-submission/530/3","title":"Multimedia Systems","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Hard Sample Mining, Data Balance, Cross-View Geo-Localization, Computer Vision","lastPublishedDoi":"10.21203/rs.3.rs-8543337/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8543337/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCross-view geo-localization (CVGL) enhances drone positioning by correlating aerial images with GPS-labeled satellite imagery.However, as data progresses through the network layers, the extracted representations become more semantically meaningful while simultaneously losing fine-grained spatial information.Moreover, the vast discrepancy in data volume between drone images and satellite images leads to uneven convergence during the training phase.Motivated by the knowledge distillation technique, we propose Confidence-Sensitive Progressive Hierarchical Distillation (CSP-HD) to enhance the effectiveness and efficiency of CVGL.Hierarchical Consistency Distillation (HCD) leverages the power of inverse self-distillation to simultaneously extract semantic features and spatial details.Progressive Adaptive Loss Weighting (PALW) dynamically adjusts the influence of HCD to alleviate short-term fluctuation during the training phase. Confidence-Sensitive Data Alignment (CSDA) dynamically adjusts the learning process based on the level of confidence to mitigate the data imbalance.Experimental results on public datasets demonstrate that our approach improves effectiveness by a margin of 1.50% to 1.84% in terms of Recall@1 and 1.50% to 1.65% in terms of average precision (AP) compared to prevailing approaches. Project page: https://1781950192.github.io/CSP-HD/}{https://1781950192.github.io/CSP-HD/.\u003c/p\u003e","manuscriptTitle":"CSP-HD: Confidence-Sensitive Progressive Hierarchical Distillation in Cross-View Geo-Localization","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-05 21:59:33","doi":"10.21203/rs.3.rs-8543337/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-02-11T13:10:33+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"108349836684400870756585683715655394626","date":"2026-02-03T05:48:41+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-03T05:43:27+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-27T08:56:08+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-10T06:55:34+00:00","index":"","fulltext":""},{"type":"submitted","content":"Multimedia Systems","date":"2026-01-07T15:34:58+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"
[email protected]","identity":"multimedia-systems","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"mmsj","sideBox":"Learn more about [Multimedia Systems](http://link.springer.com/journal/530)","snPcode":"530","submissionUrl":"https://submission.nature.com/new-submission/530/3","title":"Multimedia Systems","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"2a8958ae-7cf6-4443-9b8c-f69178e08b44","owner":[],"postedDate":"February 5th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-02-05T21:59:33+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-05 21:59:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8543337","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8543337","identity":"rs-8543337","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.