Using deep learning for defect classification on a small weld X-ray image dataset

This document provides a comparative evaluation of the performance of a deep learning network for different combinations of parameters and hyper-parameters. Although there are numerous studies that report on performance in deep learning networks for ordinary data sets, their performance on small dat...

Descripción completa

Detalles Bibliográficos
Autores: Ajmi, Chiraz, Zapata Pérez, Juan Francisco, Martínez Álvarez, José Javier, Doménech Asensi, Ginés, Ruiz Merino, Ramón Jesús
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2020
País:España
Institución:Universidad Politécnica de Cartagena(UPCT)
Repositorio:Repositorio Digital UPCT
OAI Identifier:oai:repositorio.upct.es:10317/10374
Acceso en línea:http://hdl.handle.net/10317/10374
Access Level:acceso abierto
Palabra clave:Industrial X-ray Images
Welding defects
Heterogeneities classification
Deep learning
Machine learning
Ingeniería Telemática
2203 Electrónica
id ES_bccd5e4a885b4e8fdaf6a8d22b770289
oai_identifier_str oai:repositorio.upct.es:10317/10374
network_acronym_str ES
network_name_str España
repository_id_str
spelling Using deep learning for defect classification on a small weld X-ray image datasetAjmi, ChirazZapata Pérez, Juan FranciscoMartínez Álvarez, José JavierDoménech Asensi, GinésRuiz Merino, Ramón JesúsIndustrial X-ray ImagesWelding defectsHeterogeneities classificationDeep learningMachine learningIngeniería Telemática2203 ElectrónicaThis document provides a comparative evaluation of the performance of a deep learning network for different combinations of parameters and hyper-parameters. Although there are numerous studies that report on performance in deep learning networks for ordinary data sets, their performance on small data sets is much less evaluated. The objective of this work is to demonstrate that such a challenging small data set, such as a welding X-ray image data set, can be trained and evaluated obtaining high precision and that it is possible thanks to data augmentation. In fact, this article shows that data augmentation, also a typical technique in any learning process on a large data set, plus that two image channels, such as channels B (blue) and G (green), both are replaced by the Canny edge map and a binary image provided by an adaptive Gaussian threshold, respectively, gives to the network a 3% increase in accuracy, approximately. In summary, the objective of this work is to present the methodology used and the results obtained to estimate the classification accuracy of three main classes of welding defects obtained on a small set of welding X-ray image data.The authors wants to acknowledge the work of the rest of the participants in this project, namely: J.A. López-Alcantud, P. Rubio-Ibañez, Universidad Politécnica de Cartagena, J.A. Díaz-Madrid, Centro Universitario de la Defensa - UPCT and T.J. Kazmierski, University of Southampton. This work has been partially funded by Spanish government through project numbered RTI2018-097088-B-C33 (MINECO/FEDER,UE).Springer ScienceDesarrollo de sistemas y circuitos electrónicos y microelectrónicosUniversity of TunisMinisterio de Economía y Empresa (MINECO)Fondos FEDERUnión Europea202120212020info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionapplication/pdfapplication/pdfhttp://hdl.handle.net/10317/10374reponame:Repositorio Digital UPCTinstname:Universidad Politécnica de Cartagena(UPCT)Ingléshttps://link.springer.com/article/10.1007/s10921-020-00719-9#citeasRTI2018-097088-B-C33 (MINECO/FEDER,UE).Atribución-NoComercial-CompartirIgual 3.0 España© 2020, Springer Science+Business Media, LLC, part of Springer Nature.http://creativecommons.org/licenses/by-nc-sa/3.0/es/info:eu-repo/semantics/openAccessoai:repositorio.upct.es:10317/103742026-05-15T06:39:02Z
dc.title.none.fl_str_mv Using deep learning for defect classification on a small weld X-ray image dataset
title Using deep learning for defect classification on a small weld X-ray image dataset
spellingShingle Using deep learning for defect classification on a small weld X-ray image dataset
Ajmi, Chiraz
Industrial X-ray Images
Welding defects
Heterogeneities classification
Deep learning
Machine learning
Ingeniería Telemática
2203 Electrónica
title_short Using deep learning for defect classification on a small weld X-ray image dataset
title_full Using deep learning for defect classification on a small weld X-ray image dataset
title_fullStr Using deep learning for defect classification on a small weld X-ray image dataset
title_full_unstemmed Using deep learning for defect classification on a small weld X-ray image dataset
title_sort Using deep learning for defect classification on a small weld X-ray image dataset
dc.creator.none.fl_str_mv Ajmi, Chiraz
Zapata Pérez, Juan Francisco
Martínez Álvarez, José Javier
Doménech Asensi, Ginés
Ruiz Merino, Ramón Jesús
author Ajmi, Chiraz
author_facet Ajmi, Chiraz
Zapata Pérez, Juan Francisco
Martínez Álvarez, José Javier
Doménech Asensi, Ginés
Ruiz Merino, Ramón Jesús
author_role author
author2 Zapata Pérez, Juan Francisco
Martínez Álvarez, José Javier
Doménech Asensi, Ginés
Ruiz Merino, Ramón Jesús
author2_role author
author
author
author
dc.contributor.none.fl_str_mv Desarrollo de sistemas y circuitos electrónicos y microelectrónicos
University of Tunis
Ministerio de Economía y Empresa (MINECO)
Fondos FEDER
Unión Europea
dc.subject.none.fl_str_mv Industrial X-ray Images
Welding defects
Heterogeneities classification
Deep learning
Machine learning
Ingeniería Telemática
2203 Electrónica
topic Industrial X-ray Images
Welding defects
Heterogeneities classification
Deep learning
Machine learning
Ingeniería Telemática
2203 Electrónica
description This document provides a comparative evaluation of the performance of a deep learning network for different combinations of parameters and hyper-parameters. Although there are numerous studies that report on performance in deep learning networks for ordinary data sets, their performance on small data sets is much less evaluated. The objective of this work is to demonstrate that such a challenging small data set, such as a welding X-ray image data set, can be trained and evaluated obtaining high precision and that it is possible thanks to data augmentation. In fact, this article shows that data augmentation, also a typical technique in any learning process on a large data set, plus that two image channels, such as channels B (blue) and G (green), both are replaced by the Canny edge map and a binary image provided by an adaptive Gaussian threshold, respectively, gives to the network a 3% increase in accuracy, approximately. In summary, the objective of this work is to present the methodology used and the results obtained to estimate the classification accuracy of three main classes of welding defects obtained on a small set of welding X-ray image data.
publishDate 2020
dc.date.none.fl_str_mv 2020
2021
2021
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.none.fl_str_mv http://hdl.handle.net/10317/10374
url http://hdl.handle.net/10317/10374
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.relation.none.fl_str_mv https://link.springer.com/article/10.1007/s10921-020-00719-9#citeas
RTI2018-097088-B-C33 (MINECO/FEDER,UE).
dc.rights.none.fl_str_mv Atribución-NoComercial-CompartirIgual 3.0 España
© 2020, Springer Science+Business Media, LLC, part of Springer Nature.
http://creativecommons.org/licenses/by-nc-sa/3.0/es/
info:eu-repo/semantics/openAccess
rights_invalid_str_mv Atribución-NoComercial-CompartirIgual 3.0 España
© 2020, Springer Science+Business Media, LLC, part of Springer Nature.
http://creativecommons.org/licenses/by-nc-sa/3.0/es/
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
application/pdf
dc.publisher.none.fl_str_mv Springer Science
publisher.none.fl_str_mv Springer Science
dc.source.none.fl_str_mv reponame:Repositorio Digital UPCT
instname:Universidad Politécnica de Cartagena(UPCT)
instname_str Universidad Politécnica de Cartagena(UPCT)
reponame_str Repositorio Digital UPCT
collection Repositorio Digital UPCT
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869418146276835328
score 15.300724

Similares en LA Referencia