Отрывок: Using the feedback in the neu- ral network, the initial dimension is corrected until it meets the ground truth dimension. The k-means clustering method can be used to provide feedback, which initializes the nor- malization (correction) process by taking k random boxes (centroids) as cluster heads. Clusters are then repeatedly as- signed around the nearest centroid and updated based on a certain threshold value until convergence. The proposed network’s ...
Полная запись метаданных
Поле DC Значение Язык
dc.contributor.authorLiang, T.J.-
dc.contributor.authorPan, W.G.-
dc.contributor.authorBao, H.-
dc.contributor.authorPan, F.-
dc.date.accessioned2023-02-21 10:19:26-
dc.date.available2023-02-21 10:19:26-
dc.date.issued2022-04-
dc.identifierDspace\SGAU\20230217\102066ru
dc.identifier.citationLiang TJ, Pan WG, Bao H, Pan F. Vehicle wheel weld detection based on improved YOLO v4 algorithm. Computer Optics 2022; 46(2): 271-279. DOI: 10.18287/2412-6179-CO-887.ru
dc.identifier.uri10.18287/2412-6179-CO-887-
dc.identifier.urihttp://repo.ssau.ru/handle/Zhurnal-Komputernaya-optika/Vehicle-wheel-weld-detection-based-on-improved-YOLO-v4-algorithm-102066-
dc.description.sponsorshipThe work was funded by the National Natural Science Foundation of China (Nos. 61802019, 61932012, 61871039) and the Beijing Municipal Education Commission Science and Technology Program (Nos. KM201911417009, KM201911417003, KM201911417001). Beijing Union University Research and Innovation Projects for Postgraduates (No.YZ2020K001).ru
dc.language.isoenru
dc.publisherСамарский национальный исследовательский университетru
dc.relation.ispartofseries46;2-
dc.subjectobject detection, vehicle wheel weld, YOLO v4, DIoUru
dc.titleVehicle wheel weld detection based on improved YOLO v4 algorithmru
dc.typeArticleru
dc.textpartUsing the feedback in the neu- ral network, the initial dimension is corrected until it meets the ground truth dimension. The k-means clustering method can be used to provide feedback, which initializes the nor- malization (correction) process by taking k random boxes (centroids) as cluster heads. Clusters are then repeatedly as- signed around the nearest centroid and updated based on a certain threshold value until convergence. The proposed network’s ...-
Располагается в коллекциях: Журнал "Компьютерная оптика"

Файлы этого ресурса:
Файл Описание Размер Формат  
13-Liang-Pan-Bao-Pan_KI-Lit-MA-SV-JuN2-Gr.pdfОсновная статья2.23 MBAdobe PDFПросмотреть/Открыть
Показать базовое описание ресурса Просмотр статистики
Поделиться:



Все ресурсы в архиве электронных ресурсов защищены авторским правом, все права сохранены.