Vision Foundation Models for Engineering Document Intelligence and ManufacturingInspection: A Survey

Vision Foundation Models for Engineering Document Intelligence and ManufacturingInspection: A Survey | 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 Vision Foundation Models for Engineering Document Intelligence and ManufacturingInspection: A Survey Kishore Pagidi, Anusha Gardas This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9528170/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 14 You are reading this latest preprint version Abstract Vision foundation models---large-scale architectures such as SAM, CLIP, DINOv2, and multimodal large language models (MLLMs) pretrained on internet-scale image corpora---have transformed computer vision, yet their adaptation to engineering visual data remains fragmented across disconnected research communities.We found no prior survey that jointly examines foundation models for engineering drawing understanding and manufacturing inspection, even though both domains share a common bottleneck: severe domain shift from natural-image pretraining to thin-line, symbol-dense engineering documents and controlled-lighting, texture-repetitive inspection imagery.This survey reviews 94 studies published between 2020 and 2026 across six databases, using a PRISMA-informed protocol.For engineering drawing understanding, we synthesize foundation-model methods for symbol detection, GD\&T extraction, table and BOM parsing, P\&ID graph construction, drawing-to-CAD reconstruction, and floor-plan analysis.For manufacturing inspection, we trace the progression from CLIP-based zero-shot anomaly detection through DINOv2 self-supervised methods to MLLM-based defect reasoning, distinguishing benchmark-specific performance from evidence of factory readiness.Across both domains, we identify five SAM adaptation strategies, compare adaptation regimes from frozen prompting to LoRA fine-tuning, catalogue 16+ engineering-specific datasets, and analyze six benchmark gaps, including the absence of a standardized multi-format engineering drawing benchmark.We propose DrawingBench, a modular five-track evaluation framework, and outline a research roadmap from engineering-specific pretraining corpora through hybrid foundation-model/traditional-CV deployment architectures to multimodal engineering agents with metrologically traceable perception. foundation models engineering drawings manufacturing inspection anomaly detection Segment Anything Model domain shift Full Text Additional Declarations Competing interest reported. K.R.P. is employed by SOLIDWORKS, Dassault Systèmes, and A.G. is employed by Analog Devices, Inc. The views expressed in this manuscript are the authors' own and do not necessarily represent the positions of their employers. The authors declare no other competing financial or non-financial interests. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 13 May, 2026 Reviews received at journal 12 May, 2026 Reviews received at journal 08 May, 2026 Reviews received at journal 06 May, 2026 Reviewers agreed at journal 05 May, 2026 Reviewers agreed at journal 05 May, 2026 Reviewers agreed at journal 05 May, 2026 Reviewers agreed at journal 04 May, 2026 Reviewers agreed at journal 04 May, 2026 Reviewers agreed at journal 04 May, 2026 Reviewers invited by journal 04 May, 2026 Editor assigned by journal 01 May, 2026 Submission checks completed at journal 26 Apr, 2026 First submitted to journal 25 Apr, 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-9528170","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":637173954,"identity":"320e1ca4-4049-455e-89c0-7e301c933094","order_by":0,"name":"Kishore Pagidi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABC0lEQVRIie3PMUvDQBjG8acE0uUgaxxKv8KVg5RCrF/FcpAuGTLqVihkknY9UchXyOh44eBcFHEL6CB0zaB0FepdHRTsQUeH+w/h5S4/eA/w+f5pklDQyE69xc/p7/kwObF/NPYTHkFAAErlsWTcv3+WXQHGXpZ6+3GXDoeVGQqkg1oeJpOrvGhuKZLkVXPRPGSjWof8RiBjLkJlfq7MW9KkzRmaUvXqkLCAQM2c5Kn7Jkzsye6sKqOtITs3aefSkoTGeyJnC00CQ6STTEQH85aYxW3G8VhyXuvMLEY5u3aQcTTfvHef6WgtuMJlOT2tlmoTkIvpYOVaDGYrIP577oyi/+a+9fl8Pp/tCy2FYUi7Id4hAAAAAElFTkSuQmCC","orcid":"","institution":"Northeastern University","correspondingAuthor":true,"prefix":"","firstName":"Kishore","middleName":"","lastName":"Pagidi","suffix":""},{"id":637173955,"identity":"1dea5ea6-0061-44bf-9763-6c45a03fda70","order_by":1,"name":"Anusha Gardas","email":"","orcid":"","institution":"Northeastern University","correspondingAuthor":false,"prefix":"","firstName":"Anusha","middleName":"","lastName":"Gardas","suffix":""}],"badges":[],"createdAt":"2026-04-25 20:38:58","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9528170/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9528170/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108948619,"identity":"caa1419d-289c-4ca9-a436-29bce61f7c85","added_by":"auto","created_at":"2026-05-11 06:45:20","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2583951,"visible":true,"origin":"","legend":"","description":"","filename":"ArtificialIntelligenceReviewsubmissionready.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9528170/v1_covered_6ccaa6f4-f56f-4b3c-a70f-61ba4f5f226d.pdf"}],"financialInterests":"Competing interest reported. K.R.P. is employed by SOLIDWORKS, Dassault Systèmes, and A.G. is employed by Analog Devices, Inc. The views expressed in this manuscript are the authors' own and do not necessarily represent the positions of their employers. The authors declare no other competing financial or non-financial interests.","formattedTitle":"Vision Foundation Models for Engineering Document Intelligence and ManufacturingInspection: A Survey","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":true,"email":"[email protected]","identity":"artificial-intelligence-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aire","sideBox":"Learn more about [Artificial Intelligence Review](http://link.springer.com/journal/10462)","snPcode":"10462","submissionUrl":"https://submission.nature.com/new-submission/10462/3","title":"Artificial Intelligence Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"foundation models, engineering drawings, manufacturing inspection, anomaly detection, Segment Anything Model, domain shift","lastPublishedDoi":"10.21203/rs.3.rs-9528170/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9528170/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eVision foundation models---large-scale architectures such as SAM, CLIP, DINOv2, and multimodal large language models (MLLMs) pretrained on internet-scale image corpora---have transformed computer vision, yet their adaptation to engineering visual data remains fragmented across disconnected research communities.We found no prior survey that jointly examines foundation models for engineering drawing understanding and manufacturing inspection, even though both domains share a common bottleneck: severe domain shift from natural-image pretraining to thin-line, symbol-dense engineering documents and controlled-lighting, texture-repetitive inspection imagery.This survey reviews 94 studies published between 2020 and 2026 across six databases, using a PRISMA-informed protocol.For engineering drawing understanding, we synthesize foundation-model methods for symbol detection, GD\\\u0026amp;T extraction, table and BOM parsing, P\\\u0026amp;ID graph construction, drawing-to-CAD reconstruction, and floor-plan analysis.For manufacturing inspection, we trace the progression from CLIP-based zero-shot anomaly detection through DINOv2 self-supervised methods to MLLM-based defect reasoning, distinguishing benchmark-specific performance from evidence of factory readiness.Across both domains, we identify five SAM adaptation strategies, compare adaptation regimes from frozen prompting to LoRA fine-tuning, catalogue 16+ engineering-specific datasets, and analyze six benchmark gaps, including the absence of a standardized multi-format engineering drawing benchmark.We propose DrawingBench, a modular five-track evaluation framework, and outline a research roadmap from engineering-specific pretraining corpora through hybrid foundation-model/traditional-CV deployment architectures to multimodal engineering agents with metrologically traceable perception.\u003c/p\u003e","manuscriptTitle":"Vision Foundation Models for Engineering Document Intelligence and ManufacturingInspection: A Survey","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-11 06:42:55","doi":"10.21203/rs.3.rs-9528170/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-13T10:32:17+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-12T10:12:30+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-09T02:46:27+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-06T06:03:12+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"145340608878860383094787259308311215883","date":"2026-05-06T00:59:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"103247904899032158229436359836349497725","date":"2026-05-05T14:40:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"173418353499096600302785638691989871922","date":"2026-05-05T11:32:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"59032166796734877553628591625540643297","date":"2026-05-04T19:01:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"168565401618251477899769895829958891299","date":"2026-05-04T18:33:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"132946408343077171576277316170372024465","date":"2026-05-04T16:36:20+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-04T16:34:42+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-01T13:20:58+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-27T01:47:24+00:00","index":"","fulltext":""},{"type":"submitted","content":"Artificial Intelligence Review","date":"2026-04-25T20:27:02+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"artificial-intelligence-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aire","sideBox":"Learn more about [Artificial Intelligence Review](http://link.springer.com/journal/10462)","snPcode":"10462","submissionUrl":"https://submission.nature.com/new-submission/10462/3","title":"Artificial Intelligence Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"1bab5e66-0317-40b1-99a7-038612ce1514","owner":[],"postedDate":"May 11th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-13T10:32:17+00:00","index":24,"fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-12T10:12:30+00:00","index":23,"fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-09T02:46:27+00:00","index":22,"fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-06T06:03:12+00:00","index":21,"fulltext":""},{"type":"reviewerAgreed","content":"145340608878860383094787259308311215883","date":"2026-05-06T00:59:32+00:00","index":20,"fulltext":""},{"type":"reviewerAgreed","content":"103247904899032158229436359836349497725","date":"2026-05-05T14:40:45+00:00","index":19,"fulltext":""},{"type":"reviewerAgreed","content":"173418353499096600302785638691989871922","date":"2026-05-05T11:32:56+00:00","index":18,"fulltext":""},{"type":"reviewerAgreed","content":"59032166796734877553628591625540643297","date":"2026-05-04T19:01:45+00:00","index":16,"fulltext":""},{"type":"reviewerAgreed","content":"168565401618251477899769895829958891299","date":"2026-05-04T18:33:56+00:00","index":15,"fulltext":""},{"type":"reviewerAgreed","content":"132946408343077171576277316170372024465","date":"2026-05-04T16:36:20+00:00","index":14,"fulltext":""},{"type":"reviewersInvited","content":"8","date":"2026-05-04T16:34:42+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-01T13:20:58+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-11T06:42:55+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-11 06:42:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9528170","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9528170","identity":"rs-9528170","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}