Pseudo-Supervised Defect Detection Using Robust Deep Convolutional Autoencoders

Robust Autoencoders separate the input image into a Signal(L) and a Noise(S) part which, intuitively speaking, roughly corresponds to a more stable background scene (L) and an undesired anomaly (or defect) (S). This property of the method provides a convenient theoretical basis for divorcing intermi...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Sakarya university journal of computer and information sciences Jg. 5; H. 3; S. 385 - 403
1. Verfasser: Alpdemir, Mahmut Nedim
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Sakarya University 31.12.2022
Schlagworte:
ISSN:2636-8129, 2636-8129
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Robust Autoencoders separate the input image into a Signal(L) and a Noise(S) part which, intuitively speaking, roughly corresponds to a more stable background scene (L) and an undesired anomaly (or defect) (S). This property of the method provides a convenient theoretical basis for divorcing intermittent anomalies that happen to clutter a relatively consistent background image. In this paper, we illustrate the use of Robust Deep Convolutional Autoencoders (RDCAE) for defect detection, via a pseudo-supervised training process. Our method introduces synthetic simulated defects (or structured noise) to the training process, that alleviates the scarcity of true (real-life) anomalous samples. As such, we offer a pseudo-supervised training process to devise a well-defined mechanism for deciding that the defect-normal discrimination capability of the autoencoders has reached to an acceptable point at training time. The experiment results illustrate that pseudo supervised Robust Deep Convolutional Autoencoders are very effective in identifying surface defects in an efficient way, compared to state of the art anomaly detection methods.
AbstractList Robust Autoencoders separate the input image into a Signal(L) and a Noise(S) part which, intuitively speaking, roughly corresponds to a more stable background scene (L) and an undesired anomaly (or defect) (S). This property of the method provides a convenient theoretical basis for divorcing intermittent anomalies that happen to clutter a relatively consistent background image. In this paper, we illustrate the use of Robust Deep Convolutional Autoencoders (RDCAE) for defect detection, via a pseudo-supervised training process. Our method introduces synthetic simulated defects (or structured noise) to the training process, that alleviates the scarcity of true (real-life) anomalous samples. As such, we offer a pseudo-supervised training process to devise a well-defined mechanism for deciding that the defect-normal discrimination capability of the autoencoders has reached to an acceptable point at training time. The experiment results illustrate that pseudo supervised Robust Deep Convolutional Autoencoders are very effective in identifying surface defects in an efficient way, compared to state of the art anomaly detection methods.
Author Alpdemir, Mahmut Nedim
Author_xml – sequence: 1
  givenname: Mahmut Nedim
  orcidid: 0000-0001-6411-1453
  surname: Alpdemir
  fullname: Alpdemir, Mahmut Nedim
BookMark eNpNkMtqwzAQRUVpoWmaD-jOP2BXT0tahvQVCKS0zVrI0jg4uFaQ7ED_vs6D0tWZucOcxb1D113oAKEHggsmmJSPyQ6uSUVREKJLpsgVmtCSlbkiVF__m2_RLKUdxphqTSQRE7R-TzD4kH8Oe4iHJoHPnqAG14_oRzShyzap6bbZR6iGdIxhny1CdwjtcLzaNpsPfYDOBQ8x3aOb2rYJZhdO0ebl-Wvxlq_Wr8vFfJU7SjTJuXLCCuaYrqjnSsi6Yop5VWKinGN83KwGy6Ql3mmpQFFwlOJalZVSFWNTtDx7fbA7s4_Nt40_JtjGnIIQt8bGvnEtGIxtVQtVcyc415iO755IUUnuSwaWjC5ydrkYUopQ__kINqeCzblgY8ylYPYL4_Nxxg
Cites_doi 10.1109/5.726791
10.1145/3097983.3098052
10.1145/1390156.1390294
10.13074/jent.2017.06.172255
10.1007/978-3-642-15883-4_27
10.1016/j.compind.2021.103459
10.1016/j.ijleo.2016.09.110
10.1145/2133360.2133363
10.1007/978-3-642-61717-1_13
10.1145/1541880.1541882
10.1145/3394486.3406704
10.1162/089976601750264965
10.1109/ACCESS.2020.3010274
10.1109/TKDE.2019.2947676
10.7717/peerj.453
10.1109/CVPR.2016.90
10.1007/978-3-030-32251-9_32
10.1016/j.patrec.2017.07.016
10.1016/j.sigpro.2013.12.026
10.1109/DSW.2019.8755576
10.1364/OE.27.013263
10.1023/B:AIRE.0000045502.10941.a9
10.1016/j.imavis.2011.02.002
10.1109/ICCV.2001.937541
10.3390/s20051459
10.1109/TIP.2003.819861
10.1038/nature14539
10.2478/aut-2019-0035
10.1109/ICCV.2019.00179
10.1145/1970392.1970395
10.1007/978-3-319-71249-9_3
10.1023/B:MACH.0000008084.60811.49
10.1117/12.2518459
10.1214/aoms/1177704472
10.1109/ICPR48806.2021.9412842
10.1155/2021/9948808
ContentType Journal Article
DBID AAYXX
CITATION
DOA
DOI 10.35377/saucis...1196381
DatabaseName CrossRef
DOAJ Directory of Open Access Journals
DatabaseTitle CrossRef
DatabaseTitleList
CrossRef
Database_xml – sequence: 1
  dbid: DOA
  name: DOAJ Open Access Full Text
  url: https://www.doaj.org/
  sourceTypes: Open Website
DeliveryMethod fulltext_linktorsrc
EISSN 2636-8129
EndPage 403
ExternalDocumentID oai_doaj_org_article_00abf58f4c5449028b3d175b74d63ea1
10_35377_saucis___1196381
GroupedDBID AAYXX
ALMA_UNASSIGNED_HOLDINGS
CITATION
GROUPED_DOAJ
M~E
ID FETCH-LOGICAL-c2191-48c5a53c39b2d4857fb383d86018cc34b38a9ea37a1dc978e82ec220f86b88b33
IEDL.DBID DOA
ISSN 2636-8129
IngestDate Fri Oct 03 12:52:06 EDT 2025
Sat Nov 29 05:40:48 EST 2025
IsDoiOpenAccess true
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 3
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c2191-48c5a53c39b2d4857fb383d86018cc34b38a9ea37a1dc978e82ec220f86b88b33
ORCID 0000-0001-6411-1453
OpenAccessLink https://doaj.org/article/00abf58f4c5449028b3d175b74d63ea1
PageCount 19
ParticipantIDs doaj_primary_oai_doaj_org_article_00abf58f4c5449028b3d175b74d63ea1
crossref_primary_10_35377_saucis___1196381
PublicationCentury 2000
PublicationDate 2022-12-31
PublicationDateYYYYMMDD 2022-12-31
PublicationDate_xml – month: 12
  year: 2022
  text: 2022-12-31
  day: 31
PublicationDecade 2020
PublicationTitle Sakarya university journal of computer and information sciences
PublicationYear 2022
Publisher Sakarya University
Publisher_xml – name: Sakarya University
References ref13
ref35
ref12
ref34
ref15
ref37
ref14
ref36
ref31
ref30
ref11
ref33
ref10
ref32
ref2
ref1
ref17
ref39
ref16
ref38
ref19
ref18
ref24
ref46
ref23
ref45
ref26
ref25
ref47
ref20
ref42
ref41
ref22
ref44
ref21
ref43
ref28
ref27
ref29
ref8
ref7
ref9
ref4
ref3
ref6
ref5
ref40
References_xml – ident: ref41
  doi: 10.1109/5.726791
– ident: ref11
  doi: 10.1145/3097983.3098052
– ident: ref34
  doi: 10.1145/1390156.1390294
– ident: ref16
  doi: 10.13074/jent.2017.06.172255
– ident: ref37
  doi: 10.1007/978-3-642-15883-4_27
– ident: ref43
– ident: ref45
  doi: 10.1016/j.compind.2021.103459
– ident: ref15
  doi: 10.1016/j.ijleo.2016.09.110
– ident: ref20
  doi: 10.1145/2133360.2133363
– ident: ref38
  doi: 10.1007/978-3-642-61717-1_13
– ident: ref24
– ident: ref2
  doi: 10.1145/1541880.1541882
– ident: ref7
  doi: 10.1145/3394486.3406704
– ident: ref19
  doi: 10.1162/089976601750264965
– ident: ref6
  doi: 10.1109/ACCESS.2020.3010274
– ident: ref21
  doi: 10.1109/TKDE.2019.2947676
– ident: ref47
  doi: 10.7717/peerj.453
– ident: ref5
  doi: 10.1109/CVPR.2016.90
– ident: ref9
– ident: ref10
  doi: 10.1007/978-3-030-32251-9_32
– ident: ref13
  doi: 10.1016/j.patrec.2017.07.016
– ident: ref30
– ident: ref3
  doi: 10.1016/j.sigpro.2013.12.026
– ident: ref29
  doi: 10.1109/DSW.2019.8755576
– ident: ref39
  doi: 10.1364/OE.27.013263
– ident: ref36
– ident: ref1
  doi: 10.1023/B:AIRE.0000045502.10941.a9
– ident: ref14
  doi: 10.1016/j.imavis.2011.02.002
– ident: ref32
  doi: 10.1109/ICCV.2001.937541
– ident: ref18
  doi: 10.3390/s20051459
– ident: ref46
  doi: 10.1109/TIP.2003.819861
– ident: ref28
– ident: ref4
  doi: 10.1038/nature14539
– ident: ref42
– ident: ref40
– ident: ref23
– ident: ref44
  doi: 10.2478/aut-2019-0035
– ident: ref26
  doi: 10.1109/ICCV.2019.00179
– ident: ref31
  doi: 10.1145/1970392.1970395
– ident: ref12
  doi: 10.1007/978-3-319-71249-9_3
– ident: ref22
  doi: 10.1023/B:MACH.0000008084.60811.49
– ident: ref8
– ident: ref27
  doi: 10.1117/12.2518459
– ident: ref25
  doi: 10.1214/aoms/1177704472
– ident: ref33
– ident: ref35
  doi: 10.1109/ICPR48806.2021.9412842
– ident: ref17
  doi: 10.1155/2021/9948808
SSID ssj0002991715
Score 2.2127373
Snippet Robust Autoencoders separate the input image into a Signal(L) and a Noise(S) part which, intuitively speaking, roughly corresponds to a more stable background...
SourceID doaj
crossref
SourceType Open Website
Index Database
StartPage 385
SubjectTerms anomaly detection
convolutional neural networks
defect detection
machine learning
robust autoencoders
Title Pseudo-Supervised Defect Detection Using Robust Deep Convolutional Autoencoders
URI https://doaj.org/article/00abf58f4c5449028b3d175b74d63ea1
Volume 5
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVAON
  databaseName: DOAJ Open Access Full Text
  customDbUrl:
  eissn: 2636-8129
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0002991715
  issn: 2636-8129
  databaseCode: DOA
  dateStart: 20180101
  isFulltext: true
  titleUrlDefault: https://www.doaj.org/
  providerName: Directory of Open Access Journals
– providerCode: PRVHPJ
  databaseName: ROAD: Directory of Open Access Scholarly Resources
  customDbUrl:
  eissn: 2636-8129
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0002991715
  issn: 2636-8129
  databaseCode: M~E
  dateStart: 20180101
  isFulltext: true
  titleUrlDefault: https://road.issn.org
  providerName: ISSN International Centre
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV07T8MwELZQxcCCQIAoL2VgQkqbxHbsjFBasVAqHlI3y0-pS1I1SUd-O2cnoDKxsCSKZVnR5_PdffL5M0K3OVguRHYWJ5nCMbGJi_35yZgTWExcakihTbhsgs3nfLksFjtXffmasE4euANunCRSOcod0ZQQLzWisIGQpxgxObYyEJ-EFTtkyvtgcLIpS2m3jYkpZmxcy1av6tEInIQ3uvRXINrR6w-BZXaEDvuMMLrv_uQY7dnyBL0satuaKn5r134x19ZEj9ZXXsCrCdVTZRR2-6PXSrW1b7braFKV296U_IhtU3mZSl-qfIo-ZtP3yVPc330Qa_AhQOu4ppJijQuVGcIpcwq4pOHAn7jWmMCXLKzETKZGAxO0PLM6yxLHc8UBI3yGBmVV2nMUJYoVzklsM6YJd0xC0mNprplS0kLCNkR330CIdSdxIYAaBNREh5oQokdtiB48VD8dvTp1aIA5E_2cib_m7OI_BrlEB5k_ihBEF6_QoNm09hrt622zqjc3wRzg-fw5_QK2nLu9
linkProvider Directory of Open Access Journals
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Pseudo-Supervised+Defect+Detection+Using+Robust+Deep+Convolutional+Autoencoders&rft.jtitle=Sakarya+university+journal+of+computer+and+information+sciences&rft.au=Alpdemir%2C+Mahmut+Nedim&rft.date=2022-12-31&rft.issn=2636-8129&rft.eissn=2636-8129&rft.volume=5&rft.issue=3&rft.spage=385&rft.epage=403&rft_id=info:doi/10.35377%2Fsaucis...1196381&rft.externalDBID=n%2Fa&rft.externalDocID=10_35377_saucis___1196381
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2636-8129&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2636-8129&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2636-8129&client=summon