DermFusionX: An Explainable CNN–MLP Late Fusion Framework for Multimodal Skin Lesion Classification

DermFusionX: An Explainable CNN–MLP Late Fusion Framework for Multimodal Skin Lesion Classification | 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 DermFusionX: An Explainable CNN–MLP Late Fusion Framework for Multimodal Skin Lesion Classification Vanshika Sharma This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7703674/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 Deep learning has led to extraordinary performance in domains like computer vision and natural language processing, leading to its expansion into fields such as healthcare, which demand high transparency. In clinical practice, dermatologists work with multiple data sources, such as patient metadata and lesion images, for diagnosis. Motivated by this, we propose a multimodal approach to enhance skin lesion classification on the HAM10000 dataset. We conducted extensive experiments comparing unimodal models (using only metadata or images) with multimodal models (combining both). Our evaluation included several pre-trained Convolutional Neural Networks (e.g., ResNet50, VGG19, XceptionNet, InceptionV3) and a novel custom architecture, DermFusionX. Results demonstrate that multimodal models significantly outperform their unimodal counterparts, with DermFusionX achieving precision and recall rates of 90%. To ensure transparency, we employed explainable AI techniques (LIME and SHAP) to interpret the models' decisions. Artificial Intelligence and Machine Learning Dermatology Multimodality Skin Lesion Classification Artificial Intelligence Convolutional Neural Networks Deep learning Explainability Techniques Transfer Learning Full Text Additional Declarations The authors declare no competing interests. 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-7703674","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":519977199,"identity":"0ffb7f6c-c9d1-43a9-9836-a94e9cb04049","order_by":0,"name":"Vanshika Sharma","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9UlEQVRIiWNgGAWjYJACZgaDBDBDgoHBBkgxNh4gRUsaSEsDEVoY4FoOgxl4tRgcP3v4c0FBmjwD++GDNz78OW+3tv0w0JYam2icWs7kpUnPMMgxbOBJS7ac2XY7eduZRKCWY2m5DTi0SDbkmDHzGFQwNkjwmEnzNtxONjsA1MLYcBi3lv43xp+BWuzBWv78OZdsdv4hfi38EjkG0jwGOYlgLQxsB+zMbhCwhV/ijRlQS1pyG8gvvW3JCWY3gLYk4PELG38O0GF/km37QSH244+dvdn59IcPPtTY4NSC0AulE8EqEwgpRwb2pCgeBaNgFIyCkQEAuMNcSJb7H/oAAAAASUVORK5CYII=","orcid":"https://orcid.org/0009-0006-6938-9642","institution":"Dayalbagh Educational Institute","correspondingAuthor":true,"prefix":"","firstName":"Vanshika","middleName":"","lastName":"Sharma","suffix":""}],"badges":[],"createdAt":"2025-09-24 12:26:58","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-7703674/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7703674/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92145775,"identity":"c0c056d6-a8ab-4642-9ccb-ee58b379b77f","added_by":"auto","created_at":"2025-09-25 07:07:20","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1455929,"visible":true,"origin":"","legend":"","description":"","filename":"DermFusionXAnExplainableCNNMLPLateFusionFrameworkforMultimodalSkinLesionClassification.docx","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/450a86066eba733021af08d0.docx"},{"id":92145558,"identity":"f7e05401-bd4a-4d65-bcd4-eff3aa580925","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs7703674.json","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/37921681acc56d1cff541d56.json"},{"id":92145779,"identity":"c872c5ea-5db6-42ec-96f2-349afa4992e1","added_by":"auto","created_at":"2025-09-25 07:07:20","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":123225,"visible":true,"origin":"","legend":"","description":"","filename":"rs77036740enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/1cd47dffa417d4819ee0dd50.xml"},{"id":92145776,"identity":"b7f36591-815c-4acc-9db5-ec967eadcec7","added_by":"auto","created_at":"2025-09-25 07:07:20","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":726105,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/8a3b1a365b7946edd47f4150.png"},{"id":92146684,"identity":"80ccfb5d-1b05-41b1-a444-296dfb1d4b75","added_by":"auto","created_at":"2025-09-25 07:15:20","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21838,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/c6c10c86cd0ee3a3eaac5a94.png"},{"id":92145772,"identity":"95eaaa6d-a656-4220-af2c-a4c30737514e","added_by":"auto","created_at":"2025-09-25 07:07:20","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":35765,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/5922665b6bc59fa50a489a78.png"},{"id":92145773,"identity":"51719813-9d86-42d2-8655-e73ea2eb6bde","added_by":"auto","created_at":"2025-09-25 07:07:20","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":99175,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/6728829e9f9f58be6c9ddf74.png"},{"id":92145564,"identity":"f9154346-6cdd-47b1-babd-0549a716721a","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":107251,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/2b31c84e3e2932f872e9fa39.png"},{"id":92145560,"identity":"cb270211-d93b-4045-90ce-c98903a9b1bb","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47128,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/c9b6341dfec8d1109b2d235d.jpeg"},{"id":92147063,"identity":"68134583-02e0-46eb-b70b-75d10107a7ac","added_by":"auto","created_at":"2025-09-25 07:23:21","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":205052,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/ab6b17b5d584ae7806033e53.png"},{"id":92146685,"identity":"284aca23-068e-4623-9db2-5f962a3f5be9","added_by":"auto","created_at":"2025-09-25 07:15:20","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140257,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/002b9f2fd687a9524c3f27d4.png"},{"id":92145576,"identity":"92e35a99-b8bd-4c13-805d-2dbe30daa7f7","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6439,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/23f47dc11b35e1b10c1d17e8.png"},{"id":92145568,"identity":"a349e27e-ba32-417d-a2d1-0dfd4b836630","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11584,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/1c6350d368b72ce5c7da5c54.png"},{"id":92145578,"identity":"7a41a22a-ab0c-4605-b637-cff3bfbdeb33","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":72945,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/1d91966e563e1d8337785a0b.png"},{"id":92145573,"identity":"9120ba29-fddc-43f1-8d03-7c7b4186cd6f","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5593,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/348fb1aa6b60f969bc53bb4f.png"},{"id":92146687,"identity":"73fe6f7c-7b34-419f-a897-a28de0f12bdc","added_by":"auto","created_at":"2025-09-25 07:15:20","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10447,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/e9484d74464ae3f8e14d9f07.png"},{"id":92145579,"identity":"2776be94-a9e6-4027-86cb-68d040d22efd","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12826,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/73bcfba4ee5c6cc19faa5cb6.png"},{"id":92145571,"identity":"ddbf3bf1-01a0-4443-b350-c5c68b6b3519","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19074,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/f73520c04acc7fa7ee4704b0.png"},{"id":92145569,"identity":"09598cbd-5d9e-4eba-a913-b98e52abfbfc","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47933,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/ed3777b4f7874063b2e0b528.png"},{"id":92145577,"identity":"672049b9-628e-457b-88c3-97751f3183c3","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14030,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/a1162c45aa577f03d1e5aab6.png"},{"id":92145782,"identity":"8f914638-16cd-410e-8139-a80b10a367f0","added_by":"auto","created_at":"2025-09-25 07:07:20","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":31555,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/0bb071de6af245e95d5efc6d.png"},{"id":92145582,"identity":"8d23cb25-e9ac-4573-adfc-1c893d7574cf","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1846,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/74871a72a5de8a3bc9fc9c9a.png"},{"id":92145580,"identity":"115f9d5c-5fc3-4158-a5e3-e672aedf7cee","added_by":"auto","created_at":"2025-09-25 06:59:20","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3803,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/9d7f1cfd902fd12a953c275c.png"},{"id":92145784,"identity":"353e6b53-15e0-4b31-8ab8-b35932eccc72","added_by":"auto","created_at":"2025-09-25 07:07:20","extension":"xml","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":121897,"visible":true,"origin":"","legend":"","description":"","filename":"rs77036740structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/b27457a9ac8d07612e5a767d.xml"},{"id":92145583,"identity":"a3b3ff27-03a7-4d49-8ecf-40943cb1a65a","added_by":"auto","created_at":"2025-09-25 06:59:21","extension":"html","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136375,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1/4f2c378b8ca355473e8edfce.html"},{"id":92148114,"identity":"5f717a05-f5a8-4565-b708-16ce0086b93c","added_by":"auto","created_at":"2025-09-25 07:31:22","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":692450,"visible":true,"origin":"","legend":"","description":"","filename":"DermFusionXAnExplainableCNNMLPLateFusionFrameworkforMultimodalSkinLesionClassification.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7703674/v1_covered_2f8aaa67-3ad6-4726-a0df-fe3d7c4d6e98.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eDermFusionX: An Explainable CNN–MLP Late Fusion Framework for Multimodal Skin Lesion Classification\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Dayalbagh Educational Institute","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"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":"Multimodality, Skin Lesion Classification, Artificial Intelligence, Convolutional Neural Networks, Deep learning, Explainability Techniques, Transfer Learning","lastPublishedDoi":"10.21203/rs.3.rs-7703674/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7703674/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDeep learning has led to extraordinary performance in domains like computer vision and natural language processing, leading to its expansion into fields such as healthcare, which demand high transparency. In clinical practice, dermatologists work with multiple data sources, such as patient metadata and lesion images, for diagnosis. Motivated by this, we propose a multimodal approach to enhance skin lesion classification on the HAM10000 dataset. We conducted extensive experiments comparing unimodal models (using only metadata or images) with multimodal models (combining both). Our evaluation included several pre-trained Convolutional Neural Networks (e.g., ResNet50, VGG19, XceptionNet, InceptionV3) and a novel custom architecture, DermFusionX. Results demonstrate that multimodal models significantly outperform their unimodal counterparts, with DermFusionX achieving precision and recall rates of 90%. To ensure transparency, we employed explainable AI techniques (LIME and SHAP) to interpret the models' decisions.\u003c/p\u003e","manuscriptTitle":"DermFusionX: An Explainable CNN–MLP Late Fusion Framework for Multimodal Skin Lesion Classification","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-25 06:59:15","doi":"10.21203/rs.3.rs-7703674/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":"524a908d-f7f4-4bd9-8769-cf07223b65aa","owner":[],"postedDate":"September 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":55267585,"name":"Artificial Intelligence and Machine Learning"},{"id":55267586,"name":"Dermatology"}],"tags":[],"updatedAt":"2025-09-25T06:59:15+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-25 06:59:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7703674","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7703674","identity":"rs-7703674","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}