A lightweight MobileViT with Linear Differential Attention for micro-expression recognition

A lightweight MobileViT with Linear Differential Attention for micro-expression 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 A lightweight MobileViT with Linear Differential Attention for micro-expression recognition Haiquan Wang, Kunxia Wang, Wancheng Yu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6315981/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 26 Jun, 2025 Read the published version in Signal, Image and Video Processing → Version 1 posted 5 You are reading this latest preprint version Abstract Extracting micro-expression image features using Transformer-based models is a common strategy. However, attention noise may cause the model to focus on irrelevant information. In addition, the complexity and resource consumption of the Transformer model increases significantly as the number of input tokens entered. To solve this problem, this paper proposes a Linear Differential Attention (LDA) to reduce the computation and attention noise of the MobileViT model. Firstly, We modified the self-attention computation by using piecewise functions and Gaussian kernel functions, thus reducing its complexity to linear. In this way, we obtain Linear Attention(LA). Then, we construct a pair of linear attention and use the difference between them to compute the attention score, which enhances the model's attention to key information. Finally, We use LDA to replace the Multi-Head Self-Attention in the MobileViT Block to achieve lightweight. The experimental results show that the improved MobileViT model reached 85.48% on CASME II and 76.5% on SAMM, respectively, using only 0.899G floating point operations (FLOPs) and 4.95M parameters. This demonstrates the effectiveness of our improvements. Micro-expression recognition Linear Attention Differential Operations Quadratic complexity Transformer Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 26 Jun, 2025 Read the published version in Signal, Image and Video Processing → Version 1 posted Editorial decision: Revision requested 02 Apr, 2025 Reviewers invited by journal 02 Apr, 2025 Editor assigned by journal 26 Mar, 2025 Submission checks completed at journal 26 Mar, 2025 First submitted to journal 26 Mar, 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-6315981","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":437170370,"identity":"a6fe4d7a-5e05-45c5-81bb-0d335918fb5f","order_by":0,"name":"Haiquan Wang","email":"","orcid":"","institution":"Anhui Jianzhu University","correspondingAuthor":false,"prefix":"","firstName":"Haiquan","middleName":"","lastName":"Wang","suffix":""},{"id":437170371,"identity":"293bae4b-2ce6-4238-8f43-044dc886f19c","order_by":1,"name":"Kunxia Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA50lEQVRIiWNgGAWjYBACCQhlw2DAcADEYCZaSxrpWg4DtTAQqUVyRo7h54Jf5+XNGY8/k2CosE5sYD97AK8WaZ4zxtIz+24b7mw4kCbBcCY9sYEnLwGvFjn23g3SvD23GTccOHBMgrHtcGKDBI8Bfi3MvJt/8/acs99w4GCbBOM/IrRIs/duk+b5cSBxw4HDbBKMDURokew5/82atyE5ecOBY8wWCcfSjdt4cvBrkbiRlnyb54+d7YYbxx/e+FBjLdvPfga/FjBgbANpPsDAkACk2QirB4E/QMzfQJzaUTAKRsEoGHkAAD6HSLAXa9KXAAAAAElFTkSuQmCC","orcid":"","institution":"Anhui Jianzhu University","correspondingAuthor":true,"prefix":"","firstName":"Kunxia","middleName":"","lastName":"Wang","suffix":""},{"id":437170372,"identity":"c8c99f1a-3a56-4809-84ed-00b8933bfcd5","order_by":2,"name":"Wancheng Yu","email":"","orcid":"","institution":"Anhui Jianzhu University","correspondingAuthor":false,"prefix":"","firstName":"Wancheng","middleName":"","lastName":"Yu","suffix":""}],"badges":[],"createdAt":"2025-03-27 01:38:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6315981/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6315981/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11760-025-04405-z","type":"published","date":"2025-06-26T15:57:18+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":85686198,"identity":"5b4d9280-ef50-478f-9744-eaf868663280","added_by":"auto","created_at":"2025-06-30 16:04:45","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1847045,"visible":true,"origin":"","legend":"","description":"","filename":"AlightweightMobileViTwithLinearDifferentialAttentionformicroexpressionrecognition.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6315981/v1_covered_8e5afd0d-92ff-45e0-8bf6-8ebc431ee652.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A lightweight MobileViT with Linear Differential Attention for micro-expression recognition","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Micro-expression recognition, Linear Attention, Differential Operations, Quadratic complexity, Transformer","lastPublishedDoi":"10.21203/rs.3.rs-6315981/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6315981/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eExtracting micro-expression image features using Transformer-based models is a common strategy. However, attention noise may cause the model to focus on irrelevant information. In addition, the complexity and resource consumption of the Transformer model increases significantly as the number of input tokens entered. To solve this problem, this paper proposes a Linear Differential Attention (LDA) to reduce the computation and attention noise of the MobileViT model. Firstly, We modified the self-attention computation by using piecewise functions and Gaussian kernel functions, thus reducing its complexity to linear. In this way, we obtain Linear Attention(LA). Then, we construct a pair of linear attention and use the difference between them to compute the attention score, which enhances the model's attention to key information. Finally, We use LDA to replace the Multi-Head Self-Attention in the MobileViT Block to achieve lightweight. The experimental results show that the improved MobileViT model reached 85.48% on CASME II and 76.5% on SAMM, respectively, using only 0.899G floating point operations (FLOPs) and 4.95M parameters. This demonstrates the effectiveness of our improvements.\u003c/p\u003e","manuscriptTitle":"A lightweight MobileViT with Linear Differential Attention for micro-expression recognition","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-21 01:43:09","doi":"10.21203/rs.3.rs-6315981/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-04-02T04:07:12+00:00","index":"","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-02T04:06:35+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-03-27T02:59:49+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-03-27T02:58:57+00:00","index":"","fulltext":""},{"type":"submitted","content":"Signal, Image and Video Processing","date":"2025-03-27T01:36:52+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"98fe4e0b-e4a3-4ccf-9c27-e76ad1e28b2d","owner":[],"postedDate":"April 21st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-06-30T16:02:13+00:00","versionOfRecord":{"articleIdentity":"rs-6315981","link":"https://doi.org/10.1007/s11760-025-04405-z","journal":{"identity":"signal-image-and-video-processing","isVorOnly":false,"title":"Signal, Image and Video Processing"},"publishedOn":"2025-06-26 15:57:18","publishedOnDateReadable":"June 26th, 2025"},"versionCreatedAt":"2025-04-21 01:43:09","video":"","vorDoi":"10.1007/s11760-025-04405-z","vorDoiUrl":"https://doi.org/10.1007/s11760-025-04405-z","workflowStages":[]},"version":"v1","identity":"rs-6315981","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6315981","identity":"rs-6315981","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}