Title
A Methodology for Advanced Manufacturing Defect Detection through Self-Supervised Learning on X-ray ImagesAuthor (from another institution)
xmlui.dri2xhtml.METS-1.0.item-contributorOtherinstitution
https://ror.org/05njb9z20Version
http://purl.org/coar/version/c_970fb48d4fbd8a85
Rights
© 2024 The AuthorsAccess
http://purl.org/coar/access_right/c_abf2Publisher’s version
https://doi.org/10.3390/app14072785Published at
Applied Sciences Vol. 14. N. 7. N. art. 2785Publisher
MDPIKeywords
Defect detection
manufacturing
optical quality control
deep learning ... [+]
manufacturing
optical quality control
deep learning ... [+]
Defect detection
manufacturing
optical quality control
deep learning
self-supervised learning
ODS 9 Industria, innovación e infraestructura [-]
manufacturing
optical quality control
deep learning
self-supervised learning
ODS 9 Industria, innovación e infraestructura [-]
Abstract
In industrial quality control, especially in the field of manufacturing defect detection, deep learning plays an increasingly critical role. However, the efficacy of these advanced models is often hin ... [+]
In industrial quality control, especially in the field of manufacturing defect detection, deep learning plays an increasingly critical role. However, the efficacy of these advanced models is often hindered by their need for large-scale, annotated datasets. Moreover, these datasets are mainly based on RGB images, which are very different from X-ray images. Addressing this limitation, our research proposes a methodology that incorporates domain-specific self-supervised pretraining techniques using X-ray imaging to improve defect detection capabilities in manufacturing products. We employ two pretraining approaches, SimSiam and SimMIM, to refine feature extraction from manufacturing images. The pretraining stage is carried out using an industrial dataset of 27,901 unlabeled X-ray images from a manufacturing production line. We analyze the performance of the pretraining against transfer-learning-based methods in a complex defect detection scenario using a Faster R-CNN model. We conduct evaluations on both a proprietary industrial dataset and the publicly available GDXray dataset. The findings reveal that models pretrained with domain-specific X-ray images consistently outperform those initialized with ImageNet weights. Notably, Swin Transformer models show superior results in scenarios rich in labeled data, whereas CNN backbones are more effective in limited-data environments. Moreover, we underscore the enhanced ability of the models pretrained with X-ray images in detecting critical defects, crucial for ensuring safety in industrial settings. Our study offers substantial evidence of the benefits of self-supervised learning in manufacturing defect detection, providing a solid foundation for further research and practical applications in industrial quality control. [-]
xmlui.dri2xhtml.METS-1.0.item-oaire-funderName
Gobierno VascoGobierno Vasco
xmlui.dri2xhtml.METS-1.0.item-oaire-fundingStream
Ikertalde Convocatoria 2022-2025Elkartek 2022
xmlui.dri2xhtml.METS-1.0.item-oaire-awardNumber
IT1676-22KK-2022/00049
xmlui.dri2xhtml.METS-1.0.item-oaire-awardURI
Sin informaciónSin información
xmlui.dri2xhtml.METS-1.0.item-oaire-awardTitle
Grupo de sistemas inteligentes para sistemas industrialesDeep learning REcommended Manufacturing Imperfection Novelty Detection (DREMIND)
Collections
- Articles - Engineering [639]
The following license files are associated with this item: