On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures

Defects in textured materials present a great variability, usually requiring ad-hoc solutions for each specific case. This research work proposes a solution that combines two machine learning-based approaches, convolutional autoencoders, CA; one class support vector machines, SVM. Both methods are t...

Full description

Bibliographic Details
Main Authors:	Alberto Tellaeche Iglesias, Miguel Ángel Campos Anaya, Gonzalo Pajares Martinsanz, Iker Pastor-López
Format:	Article
Language:	English
Published:	MDPI AG 2021-05-01
Series:	Sensors
Subjects:	image sensors texture inspection autoencoder SVM hybridization
Online Access:	https://www.mdpi.com/1424-8220/21/10/3339

id	doaj-9e0174a3ed8949d09419ad91d7db36a5
record_format	Article
spelling	doaj-9e0174a3ed8949d09419ad91d7db36a52021-05-31T23:44:10ZengMDPI AGSensors1424-82202021-05-01213339333910.3390/s21103339On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial TexturesAlberto Tellaeche Iglesias0Miguel Ángel Campos Anaya1Gonzalo Pajares Martinsanz2Iker Pastor-López3Computer Science, Electronics and Communication Technologies Department, University of Deusto, Avenida de las Universidades 24, 48007 Bilbao, SpainComputer Science, Electronics and Communication Technologies Department, University of Deusto, Avenida de las Universidades 24, 48007 Bilbao, SpainSoftware Engineering and Artificial Intelligence Department, Complutense University of Madrid, Calle del Prof, José García Santesmases, 9, 28040 Madrid, SpainComputer Science, Electronics and Communication Technologies Department, University of Deusto, Avenida de las Universidades 24, 48007 Bilbao, SpainDefects in textured materials present a great variability, usually requiring ad-hoc solutions for each specific case. This research work proposes a solution that combines two machine learning-based approaches, convolutional autoencoders, CA; one class support vector machines, SVM. Both methods are trained using only defect free textured images for each type of analyzed texture, labeling the samples for the SVMs in an automatic way. This work is based on two image processing streams using image sensors: (1) the CA first processes the incoming image from the input to the output, producing a reconstructed image, from which a measurement of correct or defective image is obtained; (2) the second process uses the latent layer information as input to the SVM to produce a measurement of classification. Both measurements are effectively combined, making an additional research contribution. The results obtained achieve a percentage of success of 92% on average, outperforming results of previous works.https://www.mdpi.com/1424-8220/21/10/3339image sensorstexture inspectionautoencoderSVMhybridization
collection	DOAJ
language	English
format	Article
sources	DOAJ
author	Alberto Tellaeche Iglesias Miguel Ángel Campos Anaya Gonzalo Pajares Martinsanz Iker Pastor-López
spellingShingle	Alberto Tellaeche Iglesias Miguel Ángel Campos Anaya Gonzalo Pajares Martinsanz Iker Pastor-López On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures Sensors image sensors texture inspection autoencoder SVM hybridization
author_facet	Alberto Tellaeche Iglesias Miguel Ángel Campos Anaya Gonzalo Pajares Martinsanz Iker Pastor-López
author_sort	Alberto Tellaeche Iglesias
title	On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures
title_short	On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures
title_full	On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures
title_fullStr	On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures
title_full_unstemmed	On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures
title_sort	on combining convolutional autoencoders and support vector machines for fault detection in industrial textures
publisher	MDPI AG
series	Sensors
issn	1424-8220
publishDate	2021-05-01
description	Defects in textured materials present a great variability, usually requiring ad-hoc solutions for each specific case. This research work proposes a solution that combines two machine learning-based approaches, convolutional autoencoders, CA; one class support vector machines, SVM. Both methods are trained using only defect free textured images for each type of analyzed texture, labeling the samples for the SVMs in an automatic way. This work is based on two image processing streams using image sensors: (1) the CA first processes the incoming image from the input to the output, producing a reconstructed image, from which a measurement of correct or defective image is obtained; (2) the second process uses the latent layer information as input to the SVM to produce a measurement of classification. Both measurements are effectively combined, making an additional research contribution. The results obtained achieve a percentage of success of 92% on average, outperforming results of previous works.
topic	image sensors texture inspection autoencoder SVM hybridization
url	https://www.mdpi.com/1424-8220/21/10/3339
work_keys_str_mv	AT albertotellaecheiglesias oncombiningconvolutionalautoencodersandsupportvectormachinesforfaultdetectioninindustrialtextures AT miguelangelcamposanaya oncombiningconvolutionalautoencodersandsupportvectormachinesforfaultdetectioninindustrialtextures AT gonzalopajaresmartinsanz oncombiningconvolutionalautoencodersandsupportvectormachinesforfaultdetectioninindustrialtextures AT ikerpastorlopez oncombiningconvolutionalautoencodersandsupportvectormachinesforfaultdetectioninindustrialtextures
_version_	1721416747305664512

On Combining Convolutional Autoencoders and Support Vector Machines for Fault Detection in Industrial Textures

Similar Items