Physical Privacy Protection for Human Action Recognition: A Portable, Scalable, and Secure Approach

Physical Privacy Protection for Human Action Recognition: A Portable, Scalable, and Secure Approach | 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 Physical Privacy Protection for Human Action Recognition: A Portable, Scalable, and Secure Approach Ziyi Wang, Mengyuan Liu, Peiming Li This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5631889/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 The widespread adoption of smart devices and surveillance systems has brought substantial benefits to public safety, smart homes, and health monitoring, while elevating privacy protection to a crucial concern, particularly for human action recognition technologies. Traditional video encryption methods rely on computational algorithms, which fail to fully mitigate privacy risks inherent during video capture. To tackle this issue, we propose a novel physical-layer privacy protection approach called Lens Privacy Sealing, offering an intuitive alternative to algorithmic encryption through simple hardware modifications of existing devices, enabling adjustable privacy levels to meet different recognition requirements. Conventional HAR techniques often follow a two-stage process, where human detection in encrypted videos becomes challenging due to degraded visual information, leading to significant performance loss. To balance privacy and performance, we introduce a single-stage recognition framework using a language-image pre-training model and novel transformers for spatio-temporal integration. Additionally, we develop modules to minimize interference from privacy masks, ensuring effective motion capture while preserving privacy. Extensive experiments validate that our proposed method achieves effective privacy protection while maintaining competitive recognition accuracy. Physical sciences/Mathematics and computing/Computer science Physical sciences/Mathematics and computing/Software Physical sciences/Mathematics and computing/Computational science Full Text Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryMaterials.pdf Supplementary Materials NTUEncrypted.zip NTU-Encrypted PriMo.zip PriMo 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-5631889","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":391388837,"identity":"090311f4-d06d-4fea-9adb-cca081a841b0","order_by":0,"name":"Ziyi Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/klEQVRIiWNgGAWjYDACZgaGAyCaH8JNABEGxGmRbGAmVgsMGBwgVot8O4/hYZ6aO3abz58/JsG4Iy2xgb15mwRDzR2cWhibeQwO8xx7lrztRjKbBOOZnMQGnmNlEgzHnuH2CjNQSw7b4WSzG8xALW0ViQ0SOWYSjA2HcWphA2v5dzjZuP8wVIv8G/xaeEBactsO2xkwgBzWBnSYBA9+LRLMbAWH//YdTpC4kWxskXgmzbiNJ63YIuEYbi3y/Yc3f5zx7bA9f//Bhzc+7kiW7Wc/vPHGhxrcWhgYOMCxkNgAI9lAjAQ8GhgY2B+ASHswm7EBr9JRMApGwSgYoQAAxt9Sllq1ysAAAAAASUVORK5CYII=","orcid":"https://orcid.org/0009-0006-5169-1464","institution":"Peking University","correspondingAuthor":true,"prefix":"","firstName":"Ziyi","middleName":"","lastName":"Wang","suffix":""},{"id":391388838,"identity":"04956425-a664-4b1a-817c-ff85e4476f40","order_by":1,"name":"Mengyuan Liu","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Mengyuan","middleName":"","lastName":"Liu","suffix":""},{"id":391388839,"identity":"d312158c-2059-45a8-87c9-27c21143b294","order_by":2,"name":"Peiming Li","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Peiming","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2024-12-12 13:15:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5631889/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5631889/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":71865048,"identity":"ca181746-18af-48f0-84dc-cd63d33c6d6e","added_by":"auto","created_at":"2024-12-19 09:29:08","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2824114,"visible":true,"origin":"","legend":"Article File","description":"","filename":"PhysicalPrivacyProtectionforHumanActionRecognitionAPortableScalableandSecureApproach.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5631889/v1_covered_657569cd-24fa-4142-a5b2-c88284a849a8.pdf"},{"id":71722973,"identity":"e38ce5f6-07f0-4311-a603-eb59b39f6324","added_by":"auto","created_at":"2024-12-18 05:30:42","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1241040,"visible":true,"origin":"","legend":"Supplementary Materials","description":"","filename":"SupplementaryMaterials.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5631889/v1/662db5b33e39d58175a5b825.pdf"},{"id":71722974,"identity":"9136cc0b-fe70-4e22-8b5c-b57be7dbfe93","added_by":"auto","created_at":"2024-12-18 05:30:42","extension":"zip","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1694548,"visible":true,"origin":"","legend":"NTU-Encrypted","description":"","filename":"NTUEncrypted.zip","url":"https://assets-eu.researchsquare.com/files/rs-5631889/v1/164624a78ffe0e7541aca49c.zip"},{"id":71722977,"identity":"a0f9f57f-0cfb-4a48-aa3a-5d3d9353fe3a","added_by":"auto","created_at":"2024-12-18 05:30:42","extension":"zip","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":3115276,"visible":true,"origin":"","legend":"PriMo","description":"","filename":"PriMo.zip","url":"https://assets-eu.researchsquare.com/files/rs-5631889/v1/771f87b10da449dd2672dfad.zip"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Physical Privacy Protection for Human Action Recognition: A Portable, Scalable, and Secure Approach","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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-5631889/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5631889/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The widespread adoption of smart devices and surveillance systems has brought substantial benefits to public safety, smart homes, and health monitoring, while elevating privacy protection to a crucial concern, particularly for human action recognition technologies. Traditional video encryption methods rely on computational algorithms, which fail to fully mitigate privacy risks inherent during video capture. To tackle this issue, we propose a novel physical-layer privacy protection approach called Lens Privacy Sealing, offering an intuitive alternative to algorithmic encryption through simple hardware modifications of existing devices, enabling adjustable privacy levels to meet different recognition requirements. Conventional HAR techniques often follow a two-stage process, where human detection in encrypted videos becomes challenging due to degraded visual information, leading to significant performance loss. To balance privacy and performance, we introduce a single-stage recognition framework using a language-image pre-training model and novel transformers for spatio-temporal integration. Additionally, we develop modules to minimize interference from privacy masks, ensuring effective motion capture while preserving privacy. Extensive experiments validate that our proposed method achieves effective privacy protection while maintaining competitive recognition accuracy.","manuscriptTitle":"Physical Privacy Protection for Human Action Recognition: A Portable, Scalable, and Secure Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-18 05:30:37","doi":"10.21203/rs.3.rs-5631889/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":"1f6b9a0f-962b-45c4-aa7e-d1a035383808","owner":[],"postedDate":"December 18th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":41715109,"name":"Physical sciences/Mathematics and computing/Computer science"},{"id":41715110,"name":"Physical sciences/Mathematics and computing/Software"},{"id":41715111,"name":"Physical sciences/Mathematics and computing/Computational science"}],"tags":[],"updatedAt":"2024-12-19T09:20:58+00:00","versionOfRecord":[],"versionCreatedAt":"2024-12-18 05:30:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5631889","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5631889","identity":"rs-5631889","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}