Search academic texts

Information × Registration Number 0221U104409, 0120U101924 , R & D reports Title Development of diagnostic complex based on deep neural networks for recognition of multiple surface defects of metal structures popup.stage_title Head Konovalenko Igor V., Registration Date 18-03-2021 Organization Ternopil National Technical University named after Ivan Puluj popup.description2  Number of classification models based on deep residual neural networks are constructed and their qualitative metrics on images of flat metal surfaces of rolled metal are investigated. The results showed that the proposed models with high accuracy can be used to detect and classify surface defects with high, up to 96,7% accuracy. Product Description popup.authors Baran Denys Ya. Dzyura Volodymyr O. Dozorska Oksana F. Lytvynenko Yaroslav V. Marushchak Olena V. Maruschak Pavlo O. Mykhailyshyn Roman I. Okipnyi Ihor B. Sverstyuk Andriy S. Senchyshyn Viktor S. Sorochak Andrii P. Khvostivska Liliia V. popup.nrat_date 2021-03-18 Close