A Spatiotemporal Bidirectional Mamba Network with Global–Local Skeletal Enhancement for 3D Human Pose Estimation

A Spatiotemporal Bidirectional Mamba Network with Global–Local Skeletal Enhancement for 3D Human Pose Estimation | 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 A Spatiotemporal Bidirectional Mamba Network with Global–Local Skeletal Enhancement for 3D Human Pose Estimation Chuhan Wu, Zan Wang, Guixian Zhou, Jiahao Hua, Lianke Shi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7477209/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 14 Dec, 2025 Read the published version in The Visual Computer → Version 1 posted 9 You are reading this latest preprint version Abstract 3D human pose estimation (HPE) is a cornerstone task in computer vision with diverse applications, where lifting 2D pose sequences to 3D representations has attracted significant interest. Transformer-based approaches have demonstrated robust performance but are hampered by quadratic computational complexity and insufficient bidirectional modeling capabilities. The recently introduced Mamba model mitigates these limitations through state-space models (SSMs) offering linear complexity and effective long-range dependencies; however, it falls short in modeling local skeletal interactions essential for human motion.To address this, we present BSTMamba, a bidirectional spatiotemporal SSM architecture designed specifically for monocular 3D HPE. BSTMamba integrates efficient global sequence modeling with localized convolutions and dynamic gating mechanisms to capture intricate spatiotemporal dependencies. For enhanced robustness and generalization, we introduce DisruptEnhance, a residual-compensated joint-order perturbation module that randomly disrupts joint orders at both global (full-skeleton) and local (body-part) scales, followed by feature compensation via a lightweight residual subnet. Comprehensive evaluations on the Human3.6M and MPI-INF-3DHP datasets reveal that BSTMamba attains state-of-the-art accuracy while requiring fewer parameters and lower multiply-accumulate operations (MACs) compared to prior methods. 3D Human Pose Estimation State-Space Models Bidirectional Modeling Dynamic Gating Structural Perturbation Enhancement Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 14 Dec, 2025 Read the published version in The Visual Computer → Version 1 posted Editorial decision: Revision requested 12 Oct, 2025 Reviews received at journal 01 Oct, 2025 Reviews received at journal 30 Sep, 2025 Reviewers agreed at journal 31 Aug, 2025 Reviewers agreed at journal 29 Aug, 2025 Reviewers invited by journal 28 Aug, 2025 Editor assigned by journal 28 Aug, 2025 Submission checks completed at journal 28 Aug, 2025 First submitted to journal 28 Aug, 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-7477209","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":508171047,"identity":"94538fd9-894d-453d-bda5-029976a19427","order_by":0,"name":"Chuhan Wu","email":"","orcid":"","institution":"Henan University","correspondingAuthor":false,"prefix":"","firstName":"Chuhan","middleName":"","lastName":"Wu","suffix":""},{"id":508171048,"identity":"76c81ea2-5c1e-4e8a-8714-03728f7ed84f","order_by":1,"name":"Zan Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/UlEQVRIie3NsUoDMRjA8YRAsgS7BhR9hQ8ODqRyz5IQcCxCF4eCOQLpInVtwYfo6Jhw0Fv6AAcuFcHJ4VycBG1uE+z1Rof8CfkI5MeHUCr1D6MEGy9hdveAsImnS/SRE2ZL395s8MoMJee8tmHZErz2+9cgQoUyFQdKska7HX66moAn4ZmjYnKEnNG8UXPA2+speKrHHOnpsS083wYnsKvU2vP8lCOvTD8hIrvvyPeejD77CQ8mLIEAsDISH7fQfsJK41vYSFGXc1BOq1VFs8tH0AfJhWWvrfyayZFlb7sPV6hFbV+a99viIPm1Eck4SLxgwP+OpFKpVOqvfgBK8Fs7fBfZFgAAAABJRU5ErkJggg==","orcid":"","institution":"Henan University","correspondingAuthor":true,"prefix":"","firstName":"Zan","middleName":"","lastName":"Wang","suffix":""},{"id":508171049,"identity":"c70a2593-4de9-4969-83ef-fba5dfa2031f","order_by":2,"name":"Guixian Zhou","email":"","orcid":"","institution":"Henan University","correspondingAuthor":false,"prefix":"","firstName":"Guixian","middleName":"","lastName":"Zhou","suffix":""},{"id":508171050,"identity":"66d150c7-f2e0-4de6-bf86-ca33b309b7a9","order_by":3,"name":"Jiahao Hua","email":"","orcid":"","institution":"Henan University","correspondingAuthor":false,"prefix":"","firstName":"Jiahao","middleName":"","lastName":"Hua","suffix":""},{"id":508171051,"identity":"3a5d6895-538c-41ab-b625-3e4b872a5e3a","order_by":4,"name":"Lianke Shi","email":"","orcid":"","institution":"Henan University","correspondingAuthor":false,"prefix":"","firstName":"Lianke","middleName":"","lastName":"Shi","suffix":""}],"badges":[],"createdAt":"2025-08-28 07:08:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7477209/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7477209/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00371-025-04212-0","type":"published","date":"2025-12-14T15:57:03+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":98243458,"identity":"588c13cc-dfc3-45d5-9d5d-fc6b346836ea","added_by":"auto","created_at":"2025-12-15 16:03:10","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":575351,"visible":true,"origin":"","legend":"","description":"","filename":"ArticleTitle1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7477209/v1_covered_d6573369-a9a5-4905-89c8-1192b02ffa6f.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Spatiotemporal Bidirectional Mamba Network with Global–Local Skeletal Enhancement for 3D Human Pose Estimation","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":"the-visual-computer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"tvcj","sideBox":"Learn more about [The Visual Computer](http://link.springer.com/journal/371)","snPcode":"371","submissionUrl":"https://submission.nature.com/new-submission/371/3","title":"The Visual Computer","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"3D Human Pose Estimation, State-Space Models, Bidirectional Modeling, Dynamic Gating, Structural Perturbation Enhancement","lastPublishedDoi":"10.21203/rs.3.rs-7477209/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7477209/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e3D human pose estimation (HPE) is a cornerstone task in computer vision with diverse applications, where lifting 2D pose sequences to 3D representations has attracted significant interest. Transformer-based approaches have demonstrated robust performance but are hampered by quadratic computational complexity and insufficient bidirectional modeling capabilities. The recently introduced Mamba model mitigates these limitations through state-space models (SSMs) offering linear complexity and effective long-range dependencies; however, it falls short in modeling local skeletal interactions essential for human motion.To address this, we present BSTMamba, a bidirectional spatiotemporal SSM architecture designed specifically for monocular 3D HPE. BSTMamba integrates efficient global sequence modeling with localized convolutions and dynamic gating mechanisms to capture intricate spatiotemporal dependencies. For enhanced robustness and generalization, we introduce DisruptEnhance, a residual-compensated joint-order perturbation module that randomly disrupts joint orders at both global (full-skeleton) and local (body-part) scales, followed by feature compensation via a lightweight residual subnet. Comprehensive evaluations on the Human3.6M and MPI-INF-3DHP datasets reveal that BSTMamba attains state-of-the-art accuracy while requiring fewer parameters and lower multiply-accumulate operations (MACs) compared to prior methods.\u003c/p\u003e","manuscriptTitle":"A Spatiotemporal Bidirectional Mamba Network with Global–Local Skeletal Enhancement for 3D Human Pose Estimation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-04 17:55:20","doi":"10.21203/rs.3.rs-7477209/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-12T20:29:58+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-01T11:01:36+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-30T17:31:23+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"141922403458334529595860540345233002557","date":"2025-09-01T00:45:27+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"118137936552967306740185690701864959675","date":"2025-08-30T01:40:35+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-29T03:48:31+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-28T19:29:16+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-28T10:50:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"The Visual Computer","date":"2025-08-28T06:52:40+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"the-visual-computer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"tvcj","sideBox":"Learn more about [The Visual Computer](http://link.springer.com/journal/371)","snPcode":"371","submissionUrl":"https://submission.nature.com/new-submission/371/3","title":"The Visual Computer","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"83730b1e-0ef7-4b60-a946-102a1a9c0fc7","owner":[],"postedDate":"September 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-12-15T16:00:05+00:00","versionOfRecord":{"articleIdentity":"rs-7477209","link":"https://doi.org/10.1007/s00371-025-04212-0","journal":{"identity":"the-visual-computer","isVorOnly":false,"title":"The Visual Computer"},"publishedOn":"2025-12-14 15:57:03","publishedOnDateReadable":"December 14th, 2025"},"versionCreatedAt":"2025-09-04 17:55:20","video":"","vorDoi":"10.1007/s00371-025-04212-0","vorDoiUrl":"https://doi.org/10.1007/s00371-025-04212-0","workflowStages":[]},"version":"v1","identity":"rs-7477209","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7477209","identity":"rs-7477209","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}