{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,27]],"date-time":"2026-07-27T18:46:34Z","timestamp":1785177994873,"version":"3.55.0"},"reference-count":42,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T00:00:00Z","timestamp":1760054400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:sec><jats:title>Introduction<\/jats:title><jats:p>YOLO algorithmic models are widely utilized for detecting surface defects, offering a robust and efficient approach to identifying various flaws and imperfections on material surfaces.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>In this study, we explore the integration of six distinct backbone networks within the YOLOv9 framework to optimize surface defect detection in steel strips. Specifically, we improve the YOLOv9 framework by integrating six representative backbones-ResNet50, GhostNet, MobileNetV4, FasterNet, StarNet, and RepViT-and conduct a systematic evaluation on the NEU-DET dataset and the GC10-DET dataset. Using YOLOv9-C as the baseline, we compare these backbones in terms of detection accuracy, computational complexity, and model efficiency.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Results show that RepViT achieves the best overall performance with an mAP50 of 68.8%, F1-score of 0.65, and a balanced precision-recall profile, while GhostNet offers superior computational efficiency with only 41.2 M parameters and 190.2 GFLOPs. Further validation on YOLOv5-m confirms the consistency of the results.<\/jats:p><\/jats:sec><jats:sec><jats:title>Discussion<\/jats:title><jats:p>The study offers practical guidance for backbone selection in surface defect detection tasks, highlighting the advantages of lightweight architectures for real-time industrial applications.<\/jats:p><\/jats:sec>","DOI":"10.3389\/frai.2025.1675154","type":"journal-article","created":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T05:29:51Z","timestamp":1760074191000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":6,"title":["Optimizing surface defect detection with YOLOv9: the role of advanced backbone models"],"prefix":"10.3389","volume":"8","author":[{"given":"Zhonglin","family":"Zeng","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongyang","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chi","family":"Yao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zile","family":"Dong","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shimin","family":"Cai","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,10,10]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1082","DOI":"10.1134\/S1027451018050385","article-title":"On the surface micromorphology and structure of stainless steel obtained via selective laser melting","volume":"12","author":"Afanasieva","year":"2018","journal-title":"J. Surf. Investig. X-Ray Synchrotron Neutron Tech"},{"key":"B2","doi-asserted-by":"publisher","first-page":"59","DOI":"10.7862\/rm.2025.5","article-title":"Influence of multi-pass forming on the tribological performance of aisi 430 steel sheet in deep drawing process","volume":"42","author":"Barros","year":"2025","journal-title":"Adv. Mech. Mater. Eng"},{"key":"B3","doi-asserted-by":"publisher","first-page":"164331","DOI":"10.1016\/j.ijleo.2020.164331","article-title":"Defect detection of nickel plated punched steel strip based on improved least square method","volume":"206","author":"Cao","year":"2020","journal-title":"Optik"},{"key":"B4","doi-asserted-by":"publisher","first-page":"25098","DOI":"10.1038\/s41598-025-10647-1","article-title":"Steel surface defect detection method based on improved yolov9","volume":"15","author":"Chen","year":"2025","journal-title":"Sci. Rep"},{"key":"B5","doi-asserted-by":"publisher","first-page":"12021","DOI":"10.1109\/CVPR52729.2023.01157","article-title":"\u201cRun, don't walk: chasing higher flops for faster neural networks,\u201d","author":"Chen","year":"2023","journal-title":"2023 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"B6","doi-asserted-by":"publisher","first-page":"49","DOI":"10.1007\/s00158-023-03733-3","article-title":"Enhanced gaussian-mixture-model-based nonlinear probabilistic uncertainty propagation using gaussian splitting approach","volume":"67","author":"Chen","year":"2024","journal-title":"Struct. Multidiscip. Optim"},{"key":"B7","doi-asserted-by":"publisher","first-page":"180","DOI":"10.3969\/j.issn.1671-7872.2022.02.009","article-title":"Research progress on hot-rolled strip surface defect detection based on machine vision","volume":"39","author":"Chunfeng","year":"2022","journal-title":"J. Anhui Univ. Technol"},{"key":"B8","doi-asserted-by":"publisher","first-page":"2473","DOI":"10.3390\/en18102473","article-title":"Research progress on energy-saving technologies and methods for steel metallurgy process systems\u2014a review","volume":"18","author":"Cui","year":"2025","journal-title":"Energies"},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.11929","article-title":"An image is worth 16 \u00d7 16 words: transformers for image recognition at scale","author":"Dosovitskiy","year":"2020","journal-title":"arXiv preprint arXiv:2010.11929"},{"key":"B10","doi-asserted-by":"publisher","first-page":"1096083","DOI":"10.3389\/fnbot.2023.1096083","article-title":"A multi-scale pooling convolutional neural network for accurate steel surface defects classification","volume":"17","author":"Fu","year":"2023","journal-title":"Front. Neurorobot"},{"key":"B11","doi-asserted-by":"publisher","first-page":"115438","DOI":"10.1016\/j.measurement.2024.115438","article-title":"Tsednet:task-specific encoder\u2013decoder network for surface defects of strip steel","volume":"239","author":"Guo","year":"2025","journal-title":"Measurement"},{"key":"B12","doi-asserted-by":"crossref","first-page":"1580","DOI":"10.1109\/CVPR42600.2020.00165","article-title":"\u201cGhostnet: more features from cheap operations,\u201d","volume-title":"2020 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Han","year":"2020"},{"key":"B13","doi-asserted-by":"publisher","first-page":"810876","DOI":"10.3389\/fbioe.2021.810876","article-title":"Intelligent detection of steel defects based on improved split attention networks","volume":"9","author":"Hao","year":"2022","journal-title":"Front. Bioeng. Biotechnol"},{"key":"B14","doi-asserted-by":"crossref","first-page":"770","DOI":"10.1109\/CVPR.2016.90","article-title":"\u201cDeep residual learning for image recognition,\u201d","volume-title":"2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)","author":"He","year":"2016"},{"key":"B15","doi-asserted-by":"publisher","first-page":"1493","DOI":"10.1109\/TIM.2019.2915404","article-title":"An end-to-end steel surface defect detection approach via fusing multiple hierarchical features","volume":"69","author":"He","year":"2019","journal-title":"IEEE Trans. Instrum. Meas"},{"key":"B16","doi-asserted-by":"crossref","first-page":"1314","DOI":"10.1109\/ICCV.2019.00140","article-title":"\u201cSearching for mobilenetv3,\u201d","volume-title":"2019 IEEE\/CVF International Conference on Computer Vision (ICCV)","author":"Howard","year":"2019"},{"key":"B17","doi-asserted-by":"publisher","first-page":"4571","DOI":"10.1109\/TII.2023.3327341","article-title":"Surface defect detection for no-service rails with skeleton-aware accurate and fast network","volume":"20","author":"Huang","year":"2024","journal-title":"IEEE Trans. Ind. Inform"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2410.17725","article-title":"Yolov11: an overview of the key architectural enhancements","author":"Khanam","year":"2024","journal-title":"arXiv:2410.17725"},{"key":"B19","doi-asserted-by":"publisher","first-page":"591","DOI":"10.1016\/j.jestch.2020.10.001","article-title":"Evolution of microstructure and mechanical properties of co-sic tungsten inert gas cladded coating on 304 stainless steel","volume":"24","author":"Kumar","year":"2021","journal-title":"Eng. Sci. Technol. Int. J"},{"key":"B20","doi-asserted-by":"publisher","first-page":"76","DOI":"10.1016\/j.ifacol.2018.09.412","article-title":"Real-time detection of steel strip surface defects based on improved yolo detection network","volume":"51","author":"Li","year":"2018","journal-title":"IFAC-PapersOnLine"},{"key":"B21","doi-asserted-by":"publisher","first-page":"769","DOI":"10.3390\/s25030769","article-title":"A novel yolov10-based algorithm for accurate steel surface defect detection","volume":"25","author":"Liao","year":"2025","journal-title":"Sensors"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.13462\/j.cnki.mmtamt.2020.12.007","article-title":"SVM classification of surface defect images of strip based on multi-scale lbp features","author":"Liu","year":"2020","journal-title":"Modular Mach. Tool Autom. Manuf. Technol"},{"key":"B23","doi-asserted-by":"publisher","first-page":"607","DOI":"10.3390\/biomimetics9100607","article-title":"Steel strip surface defect detection method based on improved yolov5s","volume":"9","author":"Lu","year":"2024","journal-title":"Biomimetics"},{"key":"B24","doi-asserted-by":"publisher","first-page":"1562","DOI":"10.3390\/s20061562","article-title":"Deep metallic surface defect detection: the new benchmark and detection network","volume":"20","author":"Lv","year":"2020","journal-title":"Sensors"},{"key":"B25","doi-asserted-by":"crossref","first-page":"5694","DOI":"10.1109\/CVPR52733.2024.00544","article-title":"\u201cRewrite the stars,\u201d","volume-title":"2024 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Ma","year":"2024"},{"key":"B26","doi-asserted-by":"publisher","first-page":"1119896","DOI":"10.3389\/fnbot.2023.1119896","article-title":"Research on steel rail surface defects detection based on improved yolov4 network","volume":"17","author":"Mi","year":"2023","journal-title":"Front. Neurorobot"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73661-2_5","article-title":"Mobilenetv4-universal models for the mobile ecosystem","author":"Qin","year":"2024","journal-title":"arXiv preprint arXiv:2404.10518"},{"key":"B28","doi-asserted-by":"publisher","first-page":"779","DOI":"10.1109\/CVPR.2016.91","article-title":"\u201cYou only look once: unified, real-time object detection,\u201d","author":"Redmon","year":"2016","journal-title":"2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"B29","doi-asserted-by":"publisher","first-page":"4510","DOI":"10.1109\/CVPR.2018.00474","article-title":"\u201cMobilenetv2: inverted residuals and linear bottlenecks,\u201d","author":"Sandler","year":"2018","journal-title":"2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"B30","doi-asserted-by":"publisher","first-page":"111203","DOI":"10.1016\/j.asoc.2023.111203","article-title":"A high-precision crown control strategy for hot-rolled electric steel using theoretical model-guided bo-cnn-bilstm framework","volume":"152","author":"Song","year":"2024","journal-title":"Appl. Soft Comput"},{"key":"B31","doi-asserted-by":"publisher","first-page":"858","DOI":"10.1016\/j.apsusc.2013.09.002","article-title":"A noise robust method based on completed local binary patterns for hot-rolled steel strip surface defects","volume":"285","author":"Song","year":"2013","journal-title":"Appl. Surf. Sci."},{"key":"B32","doi-asserted-by":"publisher","first-page":"6322","DOI":"10.1109\/TII.2019.2896357","article-title":"An effective method of weld defect detection and classification based on machine vision","volume":"15","author":"Sun","year":"2019","journal-title":"IEEE Trans. Ind. Inform"},{"key":"B33","doi-asserted-by":"publisher","first-page":"e38498","DOI":"10.1016\/j.heliyon.2024.e38498","article-title":"Detection and classification of surface defects on hot-rolled steel using vision transformers","volume":"10","author":"Vasan","year":"2024","journal-title":"Heliyon"},{"key":"B34","first-page":"15909","article-title":"\u201cRepvit: revisiting mobile CNN from vit perspective,\u201d","volume-title":"2024 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Wang","year":""},{"key":"B35","doi-asserted-by":"publisher","first-page":"107984","DOI":"10.48550\/arXiv.2405.14458","article-title":"Yolov10: Real-time end-to-end object detection","volume":"37","author":"Wang","year":"","journal-title":"Adv. Neural Inform. Process. Syst."},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72751-1_1","article-title":"Yolov9: Learning what you want to learn using programmable gradient information","author":"Wang","year":"2024","journal-title":"arXiv preprint arXiv:2402.13616"},{"key":"B37","doi-asserted-by":"publisher","first-page":"1042780","DOI":"10.3389\/fnbot.2022.1042780","article-title":"Enhanced you only look once x for surface defect detection of strip steel","volume":"16","author":"Wu","year":"2022","journal-title":"Front. Neurorobot"},{"key":"B38","doi-asserted-by":"publisher","first-page":"2603","DOI":"10.3390\/electronics11162603","article-title":"Surface defect detection of strip-steel based on an improved pp-yoloe-m detection network","volume":"11","author":"Zhang","year":"2022","journal-title":"Electronics"},{"key":"B39","doi-asserted-by":"publisher","first-page":"1568","DOI":"10.1111\/mice.12753","article-title":"Crack detection using fusion features-based broad learning system and image processing","volume":"36","author":"Zhang","year":"2021","journal-title":"Comput.-Aided Civ. Infrastruct. Eng"},{"key":"B40","doi-asserted-by":"publisher","first-page":"1749","DOI":"10.1007\/s00158-017-1843-9","article-title":"A novel evidence theory model dealing with correlated variables and the corresponding structural reliability analysis method","volume":"57","author":"Zhang","year":"2018","journal-title":"Struct. Multidiscip. Optim"},{"key":"B41","doi-asserted-by":"publisher","first-page":"1831","DOI":"10.1108\/EC-11-2019-0502","article-title":"Steel strip surface inspection through the combination of feature selection and multiclass classifiers","volume":"38","author":"Zhang","year":"2021","journal-title":"Eng. Comput"},{"key":"B42","doi-asserted-by":"publisher","first-page":"117393","DOI":"10.1016\/j.measurement.2025.117393","article-title":"Lwms-net: a novel defect detection network based on multi-wavelet multi-scale for steel surface defects","volume":"252","author":"Zheng","year":"2025","journal-title":"Measurement"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1675154\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T05:29:54Z","timestamp":1760074194000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1675154\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,10]]},"references-count":42,"alternative-id":["10.3389\/frai.2025.1675154"],"URL":"https:\/\/doi.org\/10.3389\/frai.2025.1675154","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10,10]]},"article-number":"1675154"}}