Exploring advanced Multi-modal Fusion Using CLIP for Autism Detection

Exploring advanced Multi-modal Fusion Using CLIP for Autism Detection | 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 Exploring advanced Multi-modal Fusion Using CLIP for Autism Detection Nidhi, Bharat Singh This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9601712/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 This study leverages the CLIP multimodal framework to enhance autism pre- diction in children by integrating questionnaire responses and image data, either individually or in combination. Unlike previous research, which often focused on a single modality, our approach aligns textual and visual information to improve diagnostic performance. Openly available autism-related datasets were used and multiple versions of the CLIP model were tested to ensure robustness between dif- ferent configurations. Training and validation used diverse data sources to improve the generalizability of the model. A comprehensive evaluation was conducted using random queries, with average accuracy measured in the top 10, 20, 30, and 60 re- trievals. The results demonstrate that the combination of questionnaire and image data offers a promising multimodal approach to autism prediction, showcasing the potential of such techniques to advance diagnostic methodologies. Artificial Intelligence and Machine Learning Clip autism image preprocessing Full Text Additional Declarations The authors declare no competing interests. The study was conducted using publicly available datasets on kaggle website, and no new data were collected from human participants. Therefore, formal ethics approval was not required. All data used in this study were anonymized and comply with the respective data usage policies. 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-9601712","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":633684046,"identity":"3c4d4eca-a996-4005-a79a-6c63b9a54ecf","order_by":0,"name":"Nidhi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIie3RsarCMBTG8a8U2iXQNYIPcUDw4mJfJaXgVHTTe5ciXNDFB1B8CSdxjGQNOItLuzgpOHo3U+mqxu2C+S8t4fzooQFcrv+YjzZE9RKOIcHqU2FFmISUVgSG3ONmTNrs9RX666JAN44a5626NPMBQlWg3Dwmnd9gRAJpslj2hVlMdcasR0j0Y0KKtc1KUtAhI0MkARmQTF6TON7riuSE6GRHvBVnFfEJ/OVXgiEXlCZzbRbTTFHAjySfkp1aN67f5o9NdevyM8spitKy/HtCalg/vRkQwPJ+6q5vzLpcLtfHdAPjekxqU+OQvAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-0991-5402","institution":"National Institute of Technology, Jamshedpur","correspondingAuthor":true,"prefix":"","firstName":"","middleName":"","lastName":"Nidhi","suffix":""},{"id":633684047,"identity":"4cc34aad-e007-47d8-9074-8bf3f01b4e96","order_by":1,"name":"Bharat Singh","email":"","orcid":"","institution":"Indian Institute of Information Technology Ranchi","correspondingAuthor":false,"prefix":"","firstName":"Bharat","middleName":"","lastName":"Singh","suffix":""}],"badges":[],"createdAt":"2026-05-03 17:31:14","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":true,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":true},"doi":"10.21203/rs.3.rs-9601712/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9601712/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108806451,"identity":"ca2db52d-68de-41b2-9d82-d9d70aea9ec2","added_by":"auto","created_at":"2026-05-08 15:28:36","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":767386,"visible":true,"origin":"","legend":"","description":"","filename":"ContributionTitle11.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9601712/v1_covered_7182330b-a635-4b92-a797-4fef4cbc1a3c.pdf"}],"financialInterests":"\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003eThe study was conducted using publicly available datasets on kaggle website, and no new data were collected from human participants. Therefore, formal ethics approval was not required. All data used in this study were anonymized and comply with the respective data usage policies.\u003c/p\u003e","formattedTitle":"\u003cp\u003eExploring advanced Multi-modal Fusion Using CLIP for Autism Detection\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"National Institue of Technology, Jamshedpur","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":"Clip, autism, image preprocessing","lastPublishedDoi":"10.21203/rs.3.rs-9601712/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9601712/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study leverages the CLIP multimodal framework to enhance autism pre- diction in children by integrating questionnaire responses and image data, either individually or in combination. Unlike previous research, which often focused on a single modality, our approach aligns textual and visual information to improve diagnostic performance. Openly available autism-related datasets were used and multiple versions of the CLIP model were tested to ensure robustness between dif- ferent configurations. Training and validation used diverse data sources to improve the generalizability of the model. A comprehensive evaluation was conducted using random queries, with average accuracy measured in the top 10, 20, 30, and 60 re- trievals. The results demonstrate that the combination of questionnaire and image data offers a promising multimodal approach to autism prediction, showcasing the potential of such techniques to advance diagnostic methodologies.\u003c/p\u003e","manuscriptTitle":"Exploring advanced Multi-modal Fusion Using CLIP for Autism Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-07 07:07:05","doi":"10.21203/rs.3.rs-9601712/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":"05bdd4b9-6192-43aa-9f45-8fa7c546bb97","owner":[],"postedDate":"May 7th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":67442944,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2026-05-07T07:07:05+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-07 07:07:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9601712","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9601712","identity":"rs-9601712","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}