GGADN: Guided Generative Adversarial Dehazing Network

GGADN: Guided Generative Adversarial Dehazing Network | 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Article GGADN: Guided Generative Adversarial Dehazing Network Zhang Jian, Wanjuan Song This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-386958/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 03 Aug, 2021 Read the published version in Soft Computing → Version 1 posted 5 You are reading this latest preprint version Abstract Image dehazing has always been a challenging topic in image processing. The development of deep learning methods, especially the Generative Adversarial Networks(GAN), provides a new way for image dehazing. In recent years, many deep learning methods based on GAN have been applied to image dehazing. However, GAN has two problems in image dehazing. Firstly, For haze image, haze not only reduces the quality of the image, but also blurs the details of the image. For Gan network, it is difficult for the generator to restore the details of the whole image while removing the haze. Secondly, GAN model is defined as a minimax problem, which weakens the loss function. It is difficult to distinguish whether GAN is making progress in the training process. Therefore, we propose a Guided Generative Adversarial Dehazing Network(GGADN). Different from other generation adversarial networks, GGADN adds a guided module on the generator. The guided module verifies the network of each layer of the generator. At the same time, the details of the map generated by each layer are strengthened. Network training is based on the pre-trained VGG feature model and L1-regularized gradient prior which is developed by new loss function parameters. From the dehazing results of synthetic images and real images, proposed method is better than the state-of-the-art dehazing methods. Theoretical Computer Science Dehazing GAN Guidance Loss function Figures Figure 1 Figure 2 Figure 3 Figure 4 Full Text Due to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the manuscript can be downloaded and accessed as a PDF. Cite Share Download PDF Status: Published Journal Publication published 03 Aug, 2021 Read the published version in Soft Computing → Version 1 posted Editorial decision: Major Revision 27 May, 2021 Editor assigned by journal 04 Apr, 2021 Reviewers invited by journal 04 Apr, 2021 Reviews received at journal 04 Apr, 2021 First submitted to journal 01 Apr, 2021 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-386958","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":20094109,"identity":"0c3bae75-457f-4ae3-a954-063a7f7753a9","order_by":0,"name":"Zhang Jian","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAvUlEQVRIiWNgGAWjYBACxmb+jw8+/KiR4ydaC3M7g7HhzJ5jxpINxGph72cwk+ZhY07ccIBYLbzNDMnGPDxsjJuPJ29g+FGxjbAWyWaGgw/nWMgwm515VsDYc+Y2YS2GzYzNBm942NjMbuQYMDO2EaHF/jAzmwTQLzzGM4jVwtjMxiYJ1CJhIEG8Fh5mUCAbSAD9cpAovzD2n2EERWV9f3vyxgc/KojQggQSDA6QpB6shVQdo2AUjIJRMEIAAK9QOila3fExAAAAAElFTkSuQmCC","orcid":"","institution":"Wuhan Sports University","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Zhang","middleName":"","lastName":"Jian","suffix":""},{"id":20094110,"identity":"49745f0a-6933-4fd7-8adb-0e1396b2bd15","order_by":1,"name":"Wanjuan Song","email":"","orcid":"","institution":"Hubei University of Education","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Wanjuan","middleName":"","lastName":"Song","suffix":""}],"badges":[],"createdAt":"2021-04-02 15:36:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-386958/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-386958/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00500-021-06049-w","type":"published","date":"2021-08-03T15:00:42+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":7768496,"identity":"b9f3f174-34b4-4696-a630-3b2b32c1bea6","added_by":"auto","created_at":"2021-04-07 20:11:51","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":93855,"visible":true,"origin":"","legend":"The structure of generator model.","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-386958/v1/9fef191cf9d54e11ed123d18.jpg"},{"id":7768253,"identity":"bfbf48dc-6e5a-4601-9afa-22166f89ba56","added_by":"auto","created_at":"2021-04-07 20:08:51","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":75771,"visible":true,"origin":"","legend":"The structure of discriminator model.","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-386958/v1/4cefc0c1841ccf92efdb0c37.jpg"},{"id":7768255,"identity":"2cd37213-f8ef-448c-b0a5-7ba1bdae13ce","added_by":"auto","created_at":"2021-04-07 20:08:51","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":203776,"visible":true,"origin":"","legend":"Qualitative comparisons with state-of-the-art dehazing methods for hazy images on synthetic data sets.","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-386958/v1/3c3eb429233b6ad1bb822546.jpg"},{"id":7768497,"identity":"0a6b54bf-3d32-4e51-8f6b-fe1355133b4a","added_by":"auto","created_at":"2021-04-07 20:11:51","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":242183,"visible":true,"origin":"","legend":"Qualitative comparisons with state-of-the-art dehazing methods for hazy images on real data sets.","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-386958/v1/9d312700ce40c3ddef469319.jpg"},{"id":13618476,"identity":"1c6f5a92-70c9-47b3-922e-f65c6ef34f5c","added_by":"auto","created_at":"2021-09-17 06:57:11","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1299755,"visible":true,"origin":"","legend":"","description":"","filename":"dehazingzhangjian20210402.pdf","url":"https://assets-eu.researchsquare.com/files/rs-386958/v1_covered.pdf"},{"id":7768970,"identity":"f484f76a-8390-4dd5-8cfe-5208769411ea","added_by":"auto","created_at":"2021-04-07 20:14:58","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":11030890,"visible":true,"origin":"","legend":"","description":"","filename":"dehazingzhangjian20210402.pdf","url":"https://assets-eu.researchsquare.com/files/rs-386958/v1_stamped.pdf"}],"financialInterests":"","formattedTitle":"GGADN: Guided Generative Adversarial Dehazing Network","fulltext":[{"header":"Full Text","content":"\u003cp\u003eDue to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the manuscript can be downloaded and accessed as a PDF.\u003c/p\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":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"soft-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"soco","sideBox":"Learn more about [Soft Computing](https://www.springer.com/journal/500)","snPcode":"500","submissionUrl":"https://submission.nature.com/new-submission/500/3","title":"Soft Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Dehazing, GAN, Guidance, Loss function","lastPublishedDoi":"10.21203/rs.3.rs-386958/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-386958/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Image dehazing has always been a challenging topic in image processing. The development of deep learning methods, especially the Generative Adversarial Networks(GAN), provides a new way for image dehazing. In recent years, many deep learning methods based on GAN have been applied to image dehazing. However, GAN has two problems in image dehazing.\n\nFirstly, For haze image, haze not only reduces the quality of the image, but also blurs the details of the image. For Gan network, it is difficult for the generator to restore the details of the whole image while removing the haze. Secondly, GAN model is defined as a minimax problem, which weakens the loss function. It is difficult to distinguish whether GAN is making progress in the training process. Therefore, we propose a Guided Generative Adversarial Dehazing Network(GGADN). Different from other generation adversarial networks, GGADN adds a guided module on the generator. The guided module verifies the network of each layer of the generator. At the same time, the details of the map generated by each layer are strengthened. Network training is based on the pre-trained VGG feature model and L1-regularized gradient prior which is developed by new loss function parameters. From the dehazing results of synthetic images and real images, proposed method is better than the state-of-the-art dehazing methods.","manuscriptTitle":"GGADN: Guided Generative Adversarial Dehazing Network","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2021-04-07 20:08:49","doi":"10.21203/rs.3.rs-386958/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major Revision","date":"2021-05-27T21:35:48+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2021-04-05T00:00:00+00:00","index":"","fulltext":""},{"type":"reviewersInvited","content":"","date":"2021-04-05T00:00:00+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2021-04-05T00:00:00+00:00","index":0,"fulltext":""},{"type":"submitted","content":"Soft Computing","date":"2021-04-01T21:14:51+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"soft-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"soco","sideBox":"Learn more about [Soft Computing](https://www.springer.com/journal/500)","snPcode":"500","submissionUrl":"https://submission.nature.com/new-submission/500/3","title":"Soft Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"1a097637-77a4-440e-bee5-8e35039e91f9","owner":[],"postedDate":"April 7th, 2021","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":3481283,"name":"Theoretical Computer Science"}],"tags":[],"updatedAt":"2021-08-22T15:06:17+00:00","versionOfRecord":{"articleIdentity":"rs-386958","link":"https://doi.org/10.1007/s00500-021-06049-w","journal":{"identity":"soft-computing","isVorOnly":false,"title":"Soft Computing"},"publishedOn":"2021-08-03 15:00:42","publishedOnDateReadable":"August 3rd, 2021"},"versionCreatedAt":"2021-04-07 20:08:49","video":"","vorDoi":"10.1007/s00500-021-06049-w","vorDoiUrl":"https://doi.org/10.1007/s00500-021-06049-w","workflowStages":[]},"version":"v1","identity":"rs-386958","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-386958","identity":"rs-386958","version":["v1"]},"buildId":"7rjqhiLT3MXkJMwkYKINL","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}