DASNet: Dual-Branch Multi-Level Attention Sheep Counting Network

DASNet: Dual-Branch Multi-Level Attention Sheep Counting 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 Article DASNet: Dual-Branch Multi-Level Attention Sheep Counting Network Chen Yini, Gao Ronghua, Li Qifeng, Zhao Hongtao, Wang Rong, Ding Luyu, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5329874/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 02 Jul, 2025 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract The grassland sheep counting task is recognized as an efficient approach for promoting the development of the livestockeconomy and maintaining the ecological balance of the grassland. In this paper, the Dual-Branch Multi-Level Attention SheepCounting Network (DASNet) is presented as a novel solution designed to address the challenges of automated sheep countingin dense grassland environments. Traditional methods have been labor-intensive and prone to inaccuracies, underscoringthe need for more efficient and accurate systems. DASNet is built on a modified VGG-19 architecture, where a dual-branchstructure is employed to integrate both shallow and deep features, thereby enhancing texture and contour detection whilereducing background noise interference. A Convolutional Block Attention Module (CBAM) is incorporated into the network tomore effectively focus on sheep regions, alongside a Multi-Level Attention Module (MAM) in the deep feature branch. TheMAM, consisting of three Light Channel and Pixel Attention Modules (LCPM), is designed to refine feature representationat both the channel and pixel levels, improving the accuracy of density map generation for sheep counting. In addition, aresidual structure is used to connect each module, facilitating feature fusion across different levels and offering increasedflexibility in handling diverse information. Experiments conducted on the self-collected Sheep1500 dataset have demonstratedthat DASNet significantly outperforms the baseline VGG-19 network, with a Mean Absolute Error (MAE) of 3.95 and a MeanSquared Error (MSE) of 4.87, compared to the baseline’s MAE of 5.39 and MSE of 6.49. DASNet is shown to be effectivein handling challenging scenarios, such as dense flocks and background noise, due to its dual-branch feature enhancementand global multi-level feature fusion. DASNet has shown promising results in terms of accuracy and computational efficiency,making it an ideal solution for practical sheep counting in precision agriculture. Physical sciences/Engineering Physical sciences/Mathematics and computing Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 02 Jul, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 24 Feb, 2025 Reviews received at journal 17 Feb, 2025 Reviewers agreed at journal 14 Feb, 2025 Reviews received at journal 21 Dec, 2024 Reviewers agreed at journal 28 Nov, 2024 Reviewers invited by journal 27 Nov, 2024 Editor assigned by journal 18 Nov, 2024 Editor invited by journal 07 Nov, 2024 Submission checks completed at journal 07 Nov, 2024 First submitted to journal 25 Oct, 2024 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-5329874","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":379839565,"identity":"dc51040d-c709-41bd-a7d4-f0904a4d851e","order_by":0,"name":"Chen Yini","email":"","orcid":"","institution":"Beijing Academy of Agricultural and Forestry Sciences","correspondingAuthor":false,"prefix":"","firstName":"Chen","middleName":"","lastName":"Yini","suffix":""},{"id":379839566,"identity":"ce0c23c1-9ea0-420f-8c78-45fa41c6a71f","order_by":1,"name":"Gao Ronghua","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2UlEQVRIiWNgGAWjYJACZoYKEMXYQIqWM2AtjcTrYWZsYyDBGnn3w8c+F867I2fefrj9wQ8Gm3x5B+ZnD/BpMTyTljx75rZnxjJnEhsbexjSLDceYDM3wKulIceYmXfb4cQZEkC/8DAcNjBs4GGTwKul/w1Qy5zD9SAtjX+I0SIvAbKl4XCCBFBLM8gWeQYCWgwkniUz8xw7bDiDJ7FxtoxBmoEBM5sZflv6kw8z89QclpdgP/7g45sKGwP59uZn+G05gMoFosP41INsaSAsMgpGwSgYBSMdAADxVEHOx0kHXAAAAABJRU5ErkJggg==","orcid":"","institution":"Beijing Academy of Agricultural and Forestry Sciences","correspondingAuthor":true,"prefix":"","firstName":"Gao","middleName":"","lastName":"Ronghua","suffix":""},{"id":379839567,"identity":"88dedd4c-b72f-4800-a19c-979b20ec8b26","order_by":2,"name":"Li Qifeng","email":"","orcid":"","institution":"Beijing Academy of Agricultural and Forestry Sciences","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Qifeng","suffix":""},{"id":379839568,"identity":"78e7ddfa-c504-46f7-a2be-c8acfb96a58e","order_by":3,"name":"Zhao Hongtao","email":"","orcid":"","institution":"North China Electric Power University","correspondingAuthor":false,"prefix":"","firstName":"Zhao","middleName":"","lastName":"Hongtao","suffix":""},{"id":379839569,"identity":"f7a68515-1fb5-4a4b-a082-72bcb84ac5d7","order_by":4,"name":"Wang Rong","email":"","orcid":"","institution":"Beijing Academy of Agricultural and Forestry Sciences","correspondingAuthor":false,"prefix":"","firstName":"Wang","middleName":"","lastName":"Rong","suffix":""},{"id":379839570,"identity":"9e4918c6-b91d-4fa1-ab74-20fa6cafbcff","order_by":5,"name":"Ding Luyu","email":"","orcid":"","institution":"Beijing Academy of Agricultural and Forestry Sciences","correspondingAuthor":false,"prefix":"","firstName":"Ding","middleName":"","lastName":"Luyu","suffix":""},{"id":379839571,"identity":"784e05cf-cb4f-4082-a56e-2f1877ec125e","order_by":6,"name":"Li Xuwen","email":"","orcid":"","institution":"Tianjin Agricultural University","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Xuwen","suffix":""}],"badges":[],"createdAt":"2024-10-25 06:08:38","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5329874/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5329874/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-97929-w","type":"published","date":"2025-07-02T15:58:45+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":86179636,"identity":"a2ef628f-a94f-49a0-b8f5-3f5ac7b6805a","added_by":"auto","created_at":"2025-07-07 16:17:59","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":17104998,"visible":true,"origin":"","legend":"","description":"","filename":"DASNetDualBranchMultiLevelAttentionSheepCountingNetwork.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5329874/v1_covered_ed0757e6-e2c8-4b26-bee6-e238b2cfbfce.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"DASNet: Dual-Branch Multi-Level Attention Sheep Counting Network","fulltext":[],"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":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-5329874/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5329874/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The grassland sheep counting task is recognized as an efficient approach for promoting the development of the livestockeconomy and maintaining the ecological balance of the grassland. In this paper, the Dual-Branch Multi-Level Attention SheepCounting Network (DASNet) is presented as a novel solution designed to address the challenges of automated sheep countingin dense grassland environments. Traditional methods have been labor-intensive and prone to inaccuracies, underscoringthe need for more efficient and accurate systems. DASNet is built on a modified VGG-19 architecture, where a dual-branchstructure is employed to integrate both shallow and deep features, thereby enhancing texture and contour detection whilereducing background noise interference. A Convolutional Block Attention Module (CBAM) is incorporated into the network tomore effectively focus on sheep regions, alongside a Multi-Level Attention Module (MAM) in the deep feature branch. TheMAM, consisting of three Light Channel and Pixel Attention Modules (LCPM), is designed to refine feature representationat both the channel and pixel levels, improving the accuracy of density map generation for sheep counting. In addition, aresidual structure is used to connect each module, facilitating feature fusion across different levels and offering increasedflexibility in handling diverse information. Experiments conducted on the self-collected Sheep1500 dataset have demonstratedthat DASNet significantly outperforms the baseline VGG-19 network, with a Mean Absolute Error (MAE) of 3.95 and a MeanSquared Error (MSE) of 4.87, compared to the baseline’s MAE of 5.39 and MSE of 6.49. DASNet is shown to be effectivein handling challenging scenarios, such as dense flocks and background noise, due to its dual-branch feature enhancementand global multi-level feature fusion. DASNet has shown promising results in terms of accuracy and computational efficiency,making it an ideal solution for practical sheep counting in precision agriculture.","manuscriptTitle":"DASNet: Dual-Branch Multi-Level Attention Sheep Counting Network","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-26 04:00:59","doi":"10.21203/rs.3.rs-5329874/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-02-24T05:53:19+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-02-17T18:50:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"24426595134822719569094415077682691405","date":"2025-02-15T03:41:37+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-12-21T10:22:59+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"250891250039657594511541826297821469784","date":"2024-11-28T05:42:24+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-11-27T18:02:33+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-11-18T11:15:48+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-11-08T04:47:58+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-11-08T04:46:27+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2024-10-25T06:05:31+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e1d35739-5585-474e-ac77-59ece692cebd","owner":[],"postedDate":"November 26th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":40447570,"name":"Physical sciences/Engineering"},{"id":40447572,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2025-07-07T16:08:49+00:00","versionOfRecord":{"articleIdentity":"rs-5329874","link":"https://doi.org/10.1038/s41598-025-97929-w","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-07-02 15:58:45","publishedOnDateReadable":"July 2nd, 2025"},"versionCreatedAt":"2024-11-26 04:00:59","video":"","vorDoi":"10.1038/s41598-025-97929-w","vorDoiUrl":"https://doi.org/10.1038/s41598-025-97929-w","workflowStages":[]},"version":"v1","identity":"rs-5329874","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5329874","identity":"rs-5329874","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}