Dual-Stream Decoupled Learning for Imbalanced Student Engagement Recognition

preprint OA: closed CC-BY-4.0

Abstract

Abstract Accurate recognition of student engagement in online learning is essential for evaluating and optimizing learning outcomes. Although extensive psychological research has demonstrated the validity of incorporating affective states into engagement assessment, most existing methods rely solely on spatiotemporal features, insufficiently representing student engagement particularly in the complex context of online learning. To bridge this gap, the proposed dual-stream architecture integrates spatiotemporal and affective features, jointly contributing to more comprehensive representations for student engagement recognition. Beyond the issue of engagement representation, the class imbalance prevalent in unconstrained online learning often leads to the severe misclassification of disengaged minority students, thereby compromising educators' trust in automated recognition. Recognizing that class imbalance primarily affects the discriminative weights of classification rather than underlying feature representations, we further decouple representation learning from classification in the training process. By freezing the dual-stream features extracted during representation learning and optimizing classification boundary through margin calibration, our decoupled learning method avoids the potential representation degradation induced by existing coupled imbalance learning methods, and is thus more effective for addressing class imbalance. Extensive experiments demonstrate that the proposed method achieves superior overall performance while significantly enhancing minority class recognition, contributing to imbalanced student engagement recognition.
Full text 13,034 characters · extracted from preprint-html · click to expand
Dual-Stream Decoupled Learning for Imbalanced Student Engagement Recognition | 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 Dual-Stream Decoupled Learning for Imbalanced Student Engagement Recognition Wenhao Liao, Yuanyuan Wang, Xinwei Zhai, Sineng Yan, Eugene Yujun Fu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9268918/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Accurate recognition of student engagement in online learning is essential for evaluating and optimizing learning outcomes. Although extensive psychological research has demonstrated the validity of incorporating affective states into engagement assessment, most existing methods rely solely on spatiotemporal features, insufficiently representing student engagement particularly in the complex context of online learning. To bridge this gap, the proposed dual-stream architecture integrates spatiotemporal and affective features, jointly contributing to more comprehensive representations for student engagement recognition. Beyond the issue of engagement representation, the class imbalance prevalent in unconstrained online learning often leads to the severe misclassification of disengaged minority students, thereby compromising educators' trust in automated recognition. Recognizing that class imbalance primarily affects the discriminative weights of classification rather than underlying feature representations, we further decouple representation learning from classification in the training process. By freezing the dual-stream features extracted during representation learning and optimizing classification boundary through margin calibration, our decoupled learning method avoids the potential representation degradation induced by existing coupled imbalance learning methods, and is thus more effective for addressing class imbalance. Extensive experiments demonstrate that the proposed method achieves superior overall performance while significantly enhancing minority class recognition, contributing to imbalanced student engagement recognition. Student engagement recognition Class imbalance Decoupled learning Affective state Online learning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 14 May, 2026 Reviewers agreed at journal 12 May, 2026 Reviewers agreed at journal 12 May, 2026 Reviewers invited by journal 16 Apr, 2026 Editor assigned by journal 07 Apr, 2026 Submission checks completed at journal 31 Mar, 2026 First submitted to journal 30 Mar, 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-9268918","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":624024406,"identity":"dea854b8-3c89-4149-9442-d562f79f584d","order_by":0,"name":"Wenhao Liao","email":"","orcid":"","institution":"Shenzhen University","correspondingAuthor":false,"prefix":"","firstName":"Wenhao","middleName":"","lastName":"Liao","suffix":""},{"id":624024408,"identity":"eb4df590-510a-4544-807a-ee94083eed6c","order_by":1,"name":"Yuanyuan Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwklEQVRIiWNgGAWjYPACGwjFQ4KWNNK1HCZBi8GNHMPPBb/O28vPSGB88LaNQd6cCC3G0jP7bic2zkhgNpzbxmC4s4GgltwN0rw9txOYJRLYpHnbGBIMDhDWsvk3b885ezaJBPbfxGrZJs3z4wBjD9AWZqK0SJ55/82atyE5cQbPw2bJOeckDDcQ0sJ3PC35Ns8fO3v59uSDH96U2cgTtEXhQgIDA2MbiMnYACQkCKgHAvl+kKF/CCscBaNgFIyCEQwAvFpATebIwqcAAAAASUVORK5CYII=","orcid":"","institution":"Shenzhen University","correspondingAuthor":true,"prefix":"","firstName":"Yuanyuan","middleName":"","lastName":"Wang","suffix":""},{"id":624024410,"identity":"5ad83020-a0db-4ee1-ad54-a81d8f645553","order_by":2,"name":"Xinwei Zhai","email":"","orcid":"","institution":"Shenzhen University","correspondingAuthor":false,"prefix":"","firstName":"Xinwei","middleName":"","lastName":"Zhai","suffix":""},{"id":624024411,"identity":"9fa5728a-6465-4542-88cc-fef93a6bcfa4","order_by":3,"name":"Sineng Yan","email":"","orcid":"","institution":"Shenzhen University","correspondingAuthor":false,"prefix":"","firstName":"Sineng","middleName":"","lastName":"Yan","suffix":""},{"id":624024412,"identity":"cf648d19-313d-412d-ab39-4616c5f56837","order_by":4,"name":"Eugene Yujun Fu","email":"","orcid":"","institution":"Education University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Eugene","middleName":"Yujun","lastName":"Fu","suffix":""}],"badges":[],"createdAt":"2026-03-30 15:10:33","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9268918/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9268918/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107708013,"identity":"5c1bcb89-4aad-4c38-a10c-1ae39651069f","added_by":"auto","created_at":"2026-04-24 09:21:39","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":13852831,"visible":true,"origin":"","legend":"","description":"","filename":"submissionv2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9268918/v1_covered_35e68f33-6bb7-4953-8d6f-b0f7e9075f8a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Dual-Stream Decoupled Learning for Imbalanced Student Engagement Recognition","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":false,"email":"","identity":"journal-on-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Journal on Image and Video Processing","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"VoR Journals","inReviewEnabled":false,"inReviewRevisionsEnabled":false},"keywords":"Student engagement recognition, Class imbalance, Decoupled learning, Affective state, Online learning","lastPublishedDoi":"10.21203/rs.3.rs-9268918/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9268918/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAccurate recognition of student engagement in online learning is essential for evaluating and optimizing learning outcomes. Although extensive psychological research has demonstrated the validity of incorporating affective states into engagement assessment, most existing methods rely solely on spatiotemporal features, insufficiently representing student engagement particularly in the complex context of online learning. To bridge this gap, the proposed dual-stream architecture integrates spatiotemporal and affective features, jointly contributing to more comprehensive representations for student engagement recognition. Beyond the issue of engagement representation, the class imbalance prevalent in unconstrained online learning often leads to the severe misclassification of disengaged minority students, thereby compromising educators' trust in automated recognition. Recognizing that class imbalance primarily affects the discriminative weights of classification rather than underlying feature representations, we further decouple representation learning from classification in the training process. By freezing the dual-stream features extracted during representation learning and optimizing classification boundary through margin calibration, our decoupled learning method avoids the potential representation degradation induced by existing coupled imbalance learning methods, and is thus more effective for addressing class imbalance. Extensive experiments demonstrate that the proposed method achieves superior overall performance while significantly enhancing minority class recognition, contributing to imbalanced student engagement recognition.\u003c/p\u003e","manuscriptTitle":"Dual-Stream Decoupled Learning for Imbalanced Student Engagement Recognition","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-24 04:57:55","doi":"10.21203/rs.3.rs-9268918/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"194617683145862508037511801382301445762","date":"2026-05-15T03:04:53+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"266247515465679215963838241423525727153","date":"2026-05-13T02:25:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"299351354086070721333917412884587881390","date":"2026-05-13T01:05:13+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-16T05:00:38+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-08T00:24:55+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-31T23:40:30+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal on Image and Video Processing","date":"2026-03-30T15:05:09+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":false,"email":"","identity":"journal-on-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Journal on Image and Video Processing","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"VoR Journals","inReviewEnabled":false,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"814d4122-1fb5-4d50-acd6-a3f44485dfc2","owner":[],"postedDate":"April 24th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"194617683145862508037511801382301445762","date":"2026-05-15T03:04:53+00:00","index":18,"fulltext":""},{"type":"reviewerAgreed","content":"266247515465679215963838241423525727153","date":"2026-05-13T02:25:48+00:00","index":17,"fulltext":""},{"type":"reviewerAgreed","content":"299351354086070721333917412884587881390","date":"2026-05-13T01:05:13+00:00","index":16,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-24T04:57:55+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-24 04:57:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9268918","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9268918","identity":"rs-9268918","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.

My notes (saved in your browser only)

Ask this paper AI returns verbatim quotes from the full text · source: preprint-html

Answers must be backed by verbatim quotes from this paper's full text. Hallucinated quotes are dropped automatically; if no verbatim passage answers the question, we say so. How this works

Citation neighborhood (no data yet)

We don't have any in-corpus citations linked to this paper yet. This is a recent paper (2026) — citers typically take a year or two to land, and the OpenAlex reference graph may still be filling in.

Source provenance

europepmc
last seen: 2026-05-20T01:45:00.602351+00:00
unpaywall
last seen: 2026-05-26T02:00:01.498150+00:00
License: CC-BY-4.0