AI-Driven Behavioural Characterisation of SeaBream Feeding Patterns in RecirculatingAquaculture Systems

AI-Driven Behavioural Characterisation of SeaBream Feeding Patterns in RecirculatingAquaculture Systems | 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 AI-Driven Behavioural Characterisation of SeaBream Feeding Patterns in RecirculatingAquaculture Systems Feliciano Domingos, Goncalo Oliveira, Isibor Kennedy Ihianle, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9393605/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 Monitoring fish behaviour in Recirculating Aquaculture Systems (RAS) is criti-cal for optimising feeding efficiency and animal welfare, yet such a task remainschallenging due to confined acoustics and intense mechanical noise. The presentresearch proposes a deep learning–based bioacoustic pipeline for classifyingbehavioural states, namely pre-feeding, feeding, post-feeding of Gilthead SeaBream (Sparus aurata) using passive acoustic monitoring; Considering the com-plexity of RAS in terms of background noise from pumps, water circulation,oxygenation systems, drum filters, oxygen sensors, air movement, human inter-vention and system interactions, the background noise is treated as a class,totalling four classes along with behavioural states. A dataset comprising multi-window segmented audio samples (1 s, 2 s, 5 s, 10 s, and 20 s) was analysedusing a structured nine-stage workflow incorporating band-limited preprocessing(50–1000 Hz), power-based signal cleansing, and Mel-spectrogram representa-tions. Across segmentation lengths, the proposed two-dimensional convolutionalneural network (2D-CNN) achieved peak performance at the 2-second window,reaching an accuracy of 99.88% and a macro F1-score of 0.9972, demonstratingstrong resilience to non-stationary hydraulic noise conditions typical of indus-trial RAS environments. Notably, high precision was consistently achieved for the pre-feeding class, substantially reducing false positive detections associated with anticipatory behaviour. The proposed framework is computationally efficient andsuitable for deployment on edge computing platforms, enabling real-time, non-invasive monitoring. These results highlight the potential of bioacoustic machinelearning (ML) and in particular deep learning (DL) approaches to support pre-cision feeding strategies, improve welfare assessment, and contribute to reducedfeed waste and environmental impact in intensive aquaculture systems. Biological sciences/Computational biology and bioinformatics Biological sciences/Ecology Earth and environmental sciences/Ecology Physical sciences/Engineering Physical sciences/Mathematics and computing Earth and environmental sciences/Ocean sciences Passive acoustic monitoring deep learning machine learning behaviour classification sea bream (Sparus aurata) recirculating aquaculture systems (RAS) Mel-spectrogram bioacoustics window segmentation feeding behaviour Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 18 May, 2026 Reviewers agreed at journal 08 May, 2026 Reviewers invited by journal 03 May, 2026 Editor assigned by journal 03 May, 2026 Editor invited by journal 28 Apr, 2026 Submission checks completed at journal 20 Apr, 2026 First submitted to journal 20 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-9393605","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":637075085,"identity":"e35accff-3fbb-42ff-a0a1-79b45e6d0f17","order_by":0,"name":"Feliciano Domingos","email":"data:image/png;base64,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","orcid":"","institution":"Nottingham Trent University","correspondingAuthor":true,"prefix":"","firstName":"Feliciano","middleName":"","lastName":"Domingos","suffix":""},{"id":637075086,"identity":"72a12dc4-9a9c-4897-a9c8-b954a6526049","order_by":1,"name":"Goncalo Oliveira","email":"","orcid":"","institution":"University of Algarve","correspondingAuthor":false,"prefix":"","firstName":"Goncalo","middleName":"","lastName":"Oliveira","suffix":""},{"id":637075087,"identity":"ba9e5244-b4d1-4225-a613-3d71d16ed293","order_by":2,"name":"Isibor Kennedy Ihianle","email":"","orcid":"","institution":"Nottingham Trent University","correspondingAuthor":false,"prefix":"","firstName":"Isibor","middleName":"Kennedy","lastName":"Ihianle","suffix":""},{"id":637075088,"identity":"abf8c482-8b59-41d3-8f79-d0b72f8c22c8","order_by":3,"name":"Joao Saraiva","email":"","orcid":"","institution":"University of Algarve","correspondingAuthor":false,"prefix":"","firstName":"Joao","middleName":"","lastName":"Saraiva","suffix":""},{"id":637075089,"identity":"e2028f2c-7838-4c0c-a5aa-10ee1cbf27ab","order_by":4,"name":"Stephanie-Carole Pieddesaux","email":"","orcid":"","institution":"Merinov","correspondingAuthor":false,"prefix":"","firstName":"Stephanie-Carole","middleName":"","lastName":"Pieddesaux","suffix":""},{"id":637075090,"identity":"ff982891-a329-462b-8cd6-e0f44f118b26","order_by":5,"name":"Ahmad Lotfi","email":"","orcid":"","institution":"Nottingham Trent University","correspondingAuthor":false,"prefix":"","firstName":"Ahmad","middleName":"","lastName":"Lotfi","suffix":""},{"id":637075091,"identity":"2e3405d5-f59c-46d0-8a3f-2a6ac39474c2","order_by":6,"name":"Pedro Machado","email":"","orcid":"","institution":"Nottingham Trent University","correspondingAuthor":false,"prefix":"","firstName":"Pedro","middleName":"","lastName":"Machado","suffix":""}],"badges":[],"createdAt":"2026-04-12 10:54:25","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9393605/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9393605/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109022185,"identity":"cf20b71e-ac8f-4685-822b-b414d41959bf","added_by":"auto","created_at":"2026-05-11 19:26:11","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1236090,"visible":true,"origin":"","legend":"","description":"","filename":"AIDrivenBehaviouralCharacterisationofSearevisedNSR.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9393605/v1_covered_f8b35411-45e8-413a-a776-54e851600c3b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"AI-Driven Behavioural Characterisation of SeaBream Feeding Patterns in RecirculatingAquaculture Systems","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":"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":"Passive acoustic monitoring, deep learning, machine learning, behaviour classification, sea bream (Sparus aurata), recirculating aquaculture systems (RAS), Mel-spectrogram, bioacoustics, window segmentation, feeding behaviour","lastPublishedDoi":"10.21203/rs.3.rs-9393605/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9393605/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Monitoring fish behaviour in Recirculating Aquaculture Systems (RAS) is criti-cal for optimising feeding efficiency and animal welfare, yet such a task remainschallenging due to confined acoustics and intense mechanical noise. The presentresearch proposes a deep learning–based bioacoustic pipeline for classifyingbehavioural states, namely pre-feeding, feeding, post-feeding of Gilthead SeaBream (Sparus aurata) using passive acoustic monitoring; Considering the com-plexity of RAS in terms of background noise from pumps, water circulation,oxygenation systems, drum filters, oxygen sensors, air movement, human inter-vention and system interactions, the background noise is treated as a class,totalling four classes along with behavioural states. A dataset comprising multi-window segmented audio samples (1 s, 2 s, 5 s, 10 s, and 20 s) was analysedusing a structured nine-stage workflow incorporating band-limited preprocessing(50–1000 Hz), power-based signal cleansing, and Mel-spectrogram representa-tions. Across segmentation lengths, the proposed two-dimensional convolutionalneural network (2D-CNN) achieved peak performance at the 2-second window,reaching an accuracy of 99.88% and a macro F1-score of 0.9972, demonstratingstrong resilience to non-stationary hydraulic noise conditions typical of indus-trial RAS environments. Notably, high precision was consistently achieved for the pre-feeding class, substantially reducing false positive detections associated with anticipatory behaviour. The proposed framework is computationally efficient andsuitable for deployment on edge computing platforms, enabling real-time, non-invasive monitoring. These results highlight the potential of bioacoustic machinelearning (ML) and in particular deep learning (DL) approaches to support pre-cision feeding strategies, improve welfare assessment, and contribute to reducedfeed waste and environmental impact in intensive aquaculture systems.","manuscriptTitle":"AI-Driven Behavioural Characterisation of SeaBream Feeding Patterns in RecirculatingAquaculture Systems","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-11 19:24:50","doi":"10.21203/rs.3.rs-9393605/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-18T19:01:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"201034426223982928133766275826358587541","date":"2026-05-08T18:33:43+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-04T02:05:13+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-04T02:04:27+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-29T03:19:35+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-20T12:24:22+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-04-20T11:35:21+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":"110039a7-f331-4317-a4df-c6f4e6a8987f","owner":[],"postedDate":"May 11th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-18T19:01:24+00:00","index":50,"fulltext":""},{"type":"reviewerAgreed","content":"201034426223982928133766275826358587541","date":"2026-05-08T18:33:43+00:00","index":48,"fulltext":""},{"type":"reviewersInvited","content":"7","date":"2026-05-04T02:05:13+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-04T02:04:27+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-29T03:19:35+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":67813728,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":67813729,"name":"Biological sciences/Ecology"},{"id":67813730,"name":"Earth and environmental sciences/Ecology"},{"id":67813731,"name":"Physical sciences/Engineering"},{"id":67813732,"name":"Physical sciences/Mathematics and computing"},{"id":67813733,"name":"Earth and environmental sciences/Ocean sciences"}],"tags":[],"updatedAt":"2026-05-11T19:24:50+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-11 19:24:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9393605","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9393605","identity":"rs-9393605","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}