From Scratch to Fine Tuning: Comparing Transfer Learning and CNN Training Strategies on Five Bangladesh-Centric Datasets

From Scratch to Fine Tuning: Comparing Transfer Learning and CNN Training Strategies on Five Bangladesh-Centric Datasets | 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 Short Report From Scratch to Fine Tuning: Comparing Transfer Learning and CNN Training Strategies on Five Bangladesh-Centric Datasets Minhaz Kamal, Md. Mushfiqul Haque, Rafid Nahiyan Farabi, Muhammad Ibrahim This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8546096/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 Convolutional neural networks (CNNs) are widely used for visual perception tasks in smart-city and agricultural settings, yet model selection in real deployments often involves practical trade-offs between performance and resource cost. In this work, we conduct a unified empirical study across five Bangladesh-centric image datasets spanning traffic monitoring, sidewalk encroachment detection, road surface condition recognition, and fine-grained agricultural variety classification. We compare three training strategies under the same dataset splits and notebook-defined training pipeline: (i) a custom CNN trained from scratch, (ii) ImageNet-pretrained ResNet50 and MobileNetV2 used as frozen feature extractors, and (iii) transfer learning by fine-tuning the same backbones. We evaluate on (a) AutoRickshaw (auto-rickshaw vs. other vehicles), (b) FootpathVisionBD (encroached vs. unencroached sidewalks), (c) RaodDamageBD (damaged vs. good road patches), (d) MangoImageBD (15 mango varieties), and (e) PaddyVisionBD (paddy variety classification). Using the values computed in the experiment notebooks, we report test accuracy and macro F1 as the primary metrics, and additionally document model parameters, model size, and training time to make the accuracy–efficiency trade-off explicit. Across datasets, fine-tuned ResNet50 provides the strongest and most consistent results on the fine-grained agricultural tasks (MangoImageBD and PaddyVisionBD), while MobileNetV2 offers a much smaller footprint with competitive performance on several smart-city tasks. Overall, the results show that the best choice is dataset-dependent: where fine-grained distinctions matter, transfer learning is usually worth the extra training cost; where memory is limited, MobileNetV2 can be a practical compromise. Artificial Intelligence and Machine Learning Convolutional neural networks Transfer learning Feature extraction ResNet50 MobileNetV2 Bangladesh 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-8546096","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Short Report","associatedPublications":[],"authors":[{"id":571164493,"identity":"7cbbc981-d269-4a98-a258-9b6ba37233c3","order_by":0,"name":"Minhaz Kamal","email":"data:image/png;base64,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","orcid":"","institution":"Department of Computer Science and Engineering, University of Dhaka","correspondingAuthor":true,"prefix":"","firstName":"Minhaz","middleName":"","lastName":"Kamal","suffix":""},{"id":571164494,"identity":"6b0abbfd-1102-4266-afdd-27bb388ab51b","order_by":1,"name":"Md. Mushfiqul Haque","email":"","orcid":"","institution":"Department of Computer Science and Engineering, University of Dhaka","correspondingAuthor":false,"prefix":"","firstName":"Md.","middleName":"Mushfiqul","lastName":"Haque","suffix":""},{"id":571164495,"identity":"4a7ac3e2-fab6-46eb-8335-48ef21e19e0a","order_by":2,"name":"Rafid Nahiyan Farabi","email":"","orcid":"","institution":"Department of Computer Science and Engineering, University of Dhaka","correspondingAuthor":false,"prefix":"","firstName":"Rafid","middleName":"Nahiyan","lastName":"Farabi","suffix":""},{"id":571164496,"identity":"ad843c3c-c671-4bd3-b6fe-b01fc9131d53","order_by":3,"name":"Muhammad Ibrahim","email":"","orcid":"","institution":"Department of Computer Science and Engineering, University of Dhaka","correspondingAuthor":false,"prefix":"","firstName":"Muhammad","middleName":"","lastName":"Ibrahim","suffix":""}],"badges":[],"createdAt":"2026-01-08 02:34:26","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-8546096/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8546096/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100357031,"identity":"99a26a2c-3c8c-4baf-976c-3e736820e88e","added_by":"auto","created_at":"2026-01-16 07:18:22","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":382963,"visible":true,"origin":"","legend":"","description":"","filename":"NeuralNetworkAssignment.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8546096/v1_covered_04109fcc-03b1-4b82-bbcb-60e9a4580aa6.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eFrom Scratch to Fine Tuning: Comparing Transfer Learning and CNN Training Strategies on Five Bangladesh-Centric Datasets\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"University of Dhaka","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":"Convolutional neural networks, Transfer learning, Feature extraction, ResNet50, MobileNetV2, Bangladesh","lastPublishedDoi":"10.21203/rs.3.rs-8546096/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8546096/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eConvolutional neural networks (CNNs) are widely used for visual perception tasks in smart-city and agricultural settings, yet model selection in real deployments often involves practical trade-offs between performance and resource cost. In this work, we conduct a unified empirical study across five Bangladesh-centric image datasets spanning traffic monitoring, sidewalk encroachment detection, road surface condition recognition, and fine-grained agricultural variety classification. We compare three training strategies under the same dataset splits and notebook-defined training pipeline: (i) a custom CNN trained from scratch, (ii) ImageNet-pretrained ResNet50 and MobileNetV2 used as frozen feature extractors, and (iii) transfer learning by fine-tuning the same backbones. We evaluate on (a) AutoRickshaw (auto-rickshaw vs. other vehicles), (b) FootpathVisionBD (encroached vs. unencroached sidewalks), (c) RaodDamageBD (damaged vs. good road patches), (d) MangoImageBD (15 mango varieties), and (e) PaddyVisionBD (paddy variety classification). Using the values computed in the experiment notebooks, we report test accuracy and macro F1 as the primary metrics, and additionally document model parameters, model size, and training time to make the accuracy–efficiency trade-off explicit. Across datasets, fine-tuned ResNet50 provides the strongest and most consistent results on the fine-grained agricultural tasks (MangoImageBD and PaddyVisionBD), while MobileNetV2 offers a much smaller footprint with competitive performance on several smart-city tasks. Overall, the results show that the best choice is dataset-dependent: where fine-grained distinctions matter, transfer learning is usually worth the extra training cost; where memory is limited, MobileNetV2 can be a practical compromise.\u003c/p\u003e","manuscriptTitle":"From Scratch to Fine Tuning: Comparing Transfer Learning and CNN Training Strategies on Five Bangladesh-Centric Datasets","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-09 05:06:39","doi":"10.21203/rs.3.rs-8546096/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":"782f4c18-0791-4eaf-97e9-042353140e47","owner":[],"postedDate":"January 9th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":60779707,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2026-01-09T05:06:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-09 05:06:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8546096","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8546096","identity":"rs-8546096","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}