LMix:Regularization Strategy for Convolutional Neural Networks

LMix:Regularization Strategy for Convolutional Neural Networks | 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 LMix:Regularization Strategy for Convolutional Neural Networks Linyu Yan, Kunpeng Zheng, Jinyao Xia, Ke Li, Hefei Ling This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-1630095/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Convolutional neural network models, as well as the training samples necessary, have grown in size in recent years. Mixed Sample Data Augmentation is provided to further improve the model's performance, and it has yielded good results.Mixed Sample Data Augmentation allows the network to generalize more effectively and improves the baseline performance of the model. The mixed sample strategies proposed so far can be broadly classified into interpolation and masking. However, interpolation-based strategies distort the data distribution, while masking-based strategies can obscure too much information. Although Mixed Sample Data Augmentation has been proven to be a viable technique for boosting deep convolutional model baseline performance, generalization ability, and robustness, there is still room for improvement in terms of image local consistency and data distribution.In this research, we present a new Mixed Sample Data Augmentation that uses random masking to increase the number of image masks while retaining the data distribution and high-frequency filtering to sharpen the images in order to emphasize recognition regions. Our experiments on CIFAR-10, CIFAR-100, Fashion-MNIST, SVHN, and Tiny-ImageNet datasets show that the LMix improves the generalization ability of state-of-the-art neural network architectures. And our method enhances the robustness of adversarial samples. Mixup Data Augmentation Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Major revision 26 Jun, 2022 Reviews received at journal 08 Jun, 2022 Reviewers agreed at journal 30 May, 2022 Reviewers invited by journal 27 May, 2022 Editor assigned by journal 27 May, 2022 Submission checks completed at journal 09 May, 2022 First submitted to journal 06 May, 2022 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-1630095","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":104617120,"identity":"8306d562-847c-4159-b79c-97a274a09263","order_by":0,"name":"Linyu Yan","email":"","orcid":"","institution":"Hubei University of Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Linyu","middleName":"","lastName":"Yan","suffix":""},{"id":104617121,"identity":"8036d8e7-77d1-41d5-aa76-6416e4b37fba","order_by":1,"name":"Kunpeng Zheng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8UlEQVRIiWNgGAWjYDACCRBhA8Q8jA0MDBUScvLEaUmDaTljYWzYQLwWIGZsq0hkOEBAh/zs5mMPvyTck+PnOdz28Os8iQTGBuaHj27g0cI451i6sUxCsbFkb2O7sew2iTx2BjZj4xw8WpglcsykJX8kJG44z9gmLblNopixgYdNGp8WNon8b9ISCQn1+8Fa5kgkNhwgoIVHIodN8kNCQoIBb2Ob5McGIrRISKSZSTMkJBjOOHOwTZrhmISxYTMBv8jPSH4G9EqCPH9POpBRUycnz9788DE+LSDAzIPCYCagHAQYf6AzRsEoGAWjYBQgAwBSMUaZw7Ti5wAAAABJRU5ErkJggg==","orcid":"","institution":"Hubei University of Technology","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Kunpeng","middleName":"","lastName":"Zheng","suffix":""},{"id":104617123,"identity":"71aa592c-28a8-4070-970c-20d1ffec9e20","order_by":2,"name":"Jinyao Xia","email":"","orcid":"","institution":"Hubei University of Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jinyao","middleName":"","lastName":"Xia","suffix":""},{"id":104617125,"identity":"8004797e-3287-461f-9050-241cdb50d2ea","order_by":3,"name":"Ke Li","email":"","orcid":"","institution":"Hubei University of Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Ke","middleName":"","lastName":"Li","suffix":""},{"id":104617128,"identity":"ede268ac-8787-4a32-8d2a-8da55bc2d4a2","order_by":4,"name":"Hefei Ling","email":"","orcid":"","institution":"Huazhong University of Science and Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Hefei","middleName":"","lastName":"Ling","suffix":""}],"badges":[],"createdAt":"2022-05-06 13:59:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-1630095/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-1630095/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":21387128,"identity":"8b2cd1ff-3e22-4927-8b49-382177aa0f17","added_by":"auto","created_at":"2022-05-12 13:30:56","extension":"pdf","order_by":7,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":374830,"visible":true,"origin":"","legend":"","description":"","filename":"LMix.pdf","url":"https://assets-eu.researchsquare.com/files/rs-1630095/v1_covered.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"LMix:Regularization Strategy for Convolutional Neural Networks","fulltext":[{"header":"Full Text","content":"This preprint is available for \u003ca href='/article/rs-1630095/latest.pdf' target='_blank'\u003edownload as a PDF\u003c/a\u003e."}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Mixup, Data Augmentation","lastPublishedDoi":"10.21203/rs.3.rs-1630095/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-1630095/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Convolutional neural network models, as well as the training samples necessary, have grown in size in recent years. Mixed Sample Data Augmentation is provided to further improve the model's performance, and it has yielded good results.Mixed Sample Data Augmentation allows the network to generalize more effectively and improves the baseline performance of the model. The mixed sample strategies proposed so far can be broadly classified into interpolation and masking. However, interpolation-based strategies distort the data distribution, while masking-based strategies can obscure too much information. Although Mixed Sample Data Augmentation has been proven to be a viable technique for boosting deep convolutional model baseline performance, generalization ability, and robustness, there is still room for improvement in terms of image local consistency and data distribution.In this research, we present a new Mixed Sample Data Augmentation that uses random masking to increase the number of image masks while retaining the data distribution and high-frequency filtering to sharpen the images in order to emphasize recognition regions. Our experiments on CIFAR-10, CIFAR-100, Fashion-MNIST, SVHN, and Tiny-ImageNet datasets show that the LMix improves the generalization ability of state-of-the-art neural network architectures. And our method enhances the robustness of adversarial samples.","manuscriptTitle":"LMix:Regularization Strategy for Convolutional Neural Networks","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-05-12 13:30:43","doi":"10.21203/rs.3.rs-1630095/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major revision","date":"2022-06-26T13:52:02+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2022-06-09T03:29:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"afb3ed3a-9bb4-4919-bc04-9f851d3739c9","date":"2022-05-30T16:45:12+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2022-05-27T10:11:22+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2022-05-27T10:09:44+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2022-05-09T13:19:36+00:00","index":"","fulltext":""},{"type":"submitted","content":"Signal, Image and Video Processing","date":"2022-05-06T13:47:02+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"49e656de-6fde-44cb-a848-f0a192137d2c","owner":[],"postedDate":"May 12th, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2022-07-28T19:14:10+00:00","versionOfRecord":[],"versionCreatedAt":"2022-05-12 13:30:43","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-1630095","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-1630095","identity":"rs-1630095","version":["v1"]},"buildId":"FbvkV6FR0MCFSLy54lSbu","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}