Swin-MFINet: Swin transformer based multi-feature integration network for detection of pixel-level surface defects

Basit Öğe Kaydı
dc.authoridUZEN, Huseyin/0000-0002-0998-2130
dc.authoridHanbay, Davut/0000-0003-2271-7865
dc.authorwosidUZEN, Huseyin/CZK-0841-2022
dc.authorwosidHanbay, Davut/AAG-8511-2019
dc.contributor.authorUzen, Huseyin
dc.contributor.authorTurkoglu, Muammer
dc.contributor.authorYanikoglu, Berrin
dc.contributor.authorHanbay, Davut
dc.date.accessioned2024-08-04T20:52:13Z
dc.date.available2024-08-04T20:52:13Z
dc.date.issued2022
dc.departmentİnönü Üniversitesien_US
dc.description.abstractAutomatic surface defect detection is critical for manufacturing industries, such as steel, fabric, and marble industries. This study proposes a Swin transformer-based model called Multi-Feature Integration Network (Swin-MFINet) for pixel-level surface defect detection. The proposed model consists of an encoder, a Swin transformer-based decoder, and Multi-Feature Integration (MFI) modules. In the encoder module of the proposed model, a pre-trained Inception network is used to extract key features from small-size datasets. In the decoder section, global semantic features are obtained from the initial features by using the Swin-transformer block, which is the newest transformer technology of today. In addition, the convolution layer is used in the last step of the decoder, since transformers are limited in acquiring small spatial details such as edges, colors, and textures, which are important in detecting some small defects. In the last module called MFI, feature maps from different decoder stages are combined, and the channel squeeze-spatial excitation block is applied to reveal important features. Finally, a prediction map is obtained by applying a convolution layer and sigmoid activation function to the MFI module output, respectively. The performance of proposed model is analyzed over MT and MVTec datasets containing surface defect images. The proposed model obtained mIoU scores of 81.37%, and 77.07% respectively, for these two datasets These results outperform the state-of-the-art for the surface defect detection problem.en_US
dc.identifier.doi10.1016/j.eswa.2022.118269
dc.identifier.issn0957-4174
dc.identifier.issn1873-6793
dc.identifier.scopus2-s2.0-85135376977en_US
dc.identifier.scopusqualityQ1en_US
dc.identifier.urihttps://doi.org/10.1016/j.eswa.2022.118269
dc.identifier.urihttps://hdl.handle.net/11616/100818
dc.identifier.volume209en_US
dc.identifier.wosWOS:000888796100009en_US
dc.identifier.wosqualityQ1en_US
dc.indekslendigikaynakWeb of Scienceen_US
dc.indekslendigikaynakScopusen_US
dc.language.isoenen_US
dc.publisherPergamon-Elsevier Science Ltden_US
dc.relation.ispartofExpert Systems With Applicationsen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/closedAccessen_US
dc.subjectPixel-Level Surface Defects Detectionen_US
dc.subjectSwin Transformersen_US
dc.subjectEncoder-Decoder Networken_US
dc.subjectConvolutional Neural Networken_US
dc.titleSwin-MFINet: Swin transformer based multi-feature integration network for detection of pixel-level surface defectsen_US
dc.typeArticleen_US

Dosyalar

Koleksiyon

WoS İndeksli Yayınlar Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu