Convolutional autoencoder based on latent subspace projection for anomaly detection

Image anomaly detection (AD) is widely researched on many occasions in computer vision tasks. High-dimensional data, such as image data, with noise and complex background is still challenging to detect anomalies under the situation that imbalanced or incomplete data are available. Some deep learning...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Methods (San Diego, Calif.) Ročník 214; s. 48 - 59
Hlavní autoři: Yu, Qien, Li, Chen, Zhu, Ye, Kurita, Takio
Médium: Journal Article
Jazyk:angličtina
Vydáno: United States Elsevier Inc 01.06.2023
Témata:
ISSN:1046-2023, 1095-9130, 1095-9130
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Image anomaly detection (AD) is widely researched on many occasions in computer vision tasks. High-dimensional data, such as image data, with noise and complex background is still challenging to detect anomalies under the situation that imbalanced or incomplete data are available. Some deep learning methods can be trained in an unsupervised way and map the original input into low-dimensional manifolds to predict larger differences in anomalies according to normal ones by dimension reduction. However, training a single low-dimension latent space is limited to present the low-dimensional features due to the fact that the noise and irreverent features are mapped into this space, resulting in that the manifolds are not discriminative for detecting anomalies. To address this problem, a new autoencoder framework is proposed in this study with two trainable mutually orthogonal complementary subspaces in the latent space, by latent subspace projection (LSP) mechanism, which is named as LSP-CAE. Specifically, latent subspace projection is used to train the latent image subspace (LIS) and the latent kernel subspace (LKS) in the latent space of the autoencoder-like model respectively, which can enhance learning power of different features from the input instance. The features of normal data are projected into the latent image subspace, while the latent kernel subspace is trained to extract the irrelevant information according to normal features by end-to-end training. To verify the generality and effectiveness of the proposed method, we replace the convolutional network with the fully-connected network contucted in the real-world medical datasets. The anomaly score based on projection norms in two subspace is used to evaluate the anomalies in the testing. Consequently, our proposed method can achieve the best performance according to four public datasets in comparison of the state-of-the-art methods. •We propose a latent subspace projection (LSP) constraints by incorporating deep encoder-decoder structure.•Complementary orthogonal subspaces can be trained in the end-to-end fashion to learn discriminative latent manifolds.•The proposed method shows the best performance in comparison of the state-of-the-art methods on public datasets.
AbstractList Image anomaly detection (AD) is widely researched on many occasions in computer vision tasks. High-dimensional data, such as image data, with noise and complex background is still challenging to detect anomalies under the situation that imbalanced or incomplete data are available. Some deep learning methods can be trained in an unsupervised way and map the original input into low-dimensional manifolds to predict larger differences in anomalies according to normal ones by dimension reduction. However, training a single low-dimension latent space is limited to present the low-dimensional features due to the fact that the noise and irreverent features are mapped into this space, resulting in that the manifolds are not discriminative for detecting anomalies. To address this problem, a new autoencoder framework is proposed in this study with two trainable mutually orthogonal complementary subspaces in the latent space, by latent subspace projection (LSP) mechanism, which is named as LSP-CAE. Specifically, latent subspace projection is used to train the latent image subspace (LIS) and the latent kernel subspace (LKS) in the latent space of the autoencoder-like model respectively, which can enhance learning power of different features from the input instance. The features of normal data are projected into the latent image subspace, while the latent kernel subspace is trained to extract the irrelevant information according to normal features by end-to-end training. To verify the generality and effectiveness of the proposed method, we replace the convolutional network with the fully-connected network contucted in the real-world medical datasets. The anomaly score based on projection norms in two subspace is used to evaluate the anomalies in the testing. Consequently, our proposed method can achieve the best performance according to four public datasets in comparison of the state-of-the-art methods.
Image anomaly detection (AD) is widely researched on many occasions in computer vision tasks. High-dimensional data, such as image data, with noise and complex background is still challenging to detect anomalies under the situation that imbalanced or incomplete data are available. Some deep learning methods can be trained in an unsupervised way and map the original input into low-dimensional manifolds to predict larger differences in anomalies according to normal ones by dimension reduction. However, training a single low-dimension latent space is limited to present the low-dimensional features due to the fact that the noise and irreverent features are mapped into this space, resulting in that the manifolds are not discriminative for detecting anomalies. To address this problem, a new autoencoder framework is proposed in this study with two trainable mutually orthogonal complementary subspaces in the latent space, by latent subspace projection (LSP) mechanism, which is named as LSP-CAE. Specifically, latent subspace projection is used to train the latent image subspace (LIS) and the latent kernel subspace (LKS) in the latent space of the autoencoder-like model respectively, which can enhance learning power of different features from the input instance. The features of normal data are projected into the latent image subspace, while the latent kernel subspace is trained to extract the irrelevant information according to normal features by end-to-end training. To verify the generality and effectiveness of the proposed method, we replace the convolutional network with the fully-connected network contucted in the real-world medical datasets. The anomaly score based on projection norms in two subspace is used to evaluate the anomalies in the testing. Consequently, our proposed method can achieve the best performance according to four public datasets in comparison of the state-of-the-art methods.Image anomaly detection (AD) is widely researched on many occasions in computer vision tasks. High-dimensional data, such as image data, with noise and complex background is still challenging to detect anomalies under the situation that imbalanced or incomplete data are available. Some deep learning methods can be trained in an unsupervised way and map the original input into low-dimensional manifolds to predict larger differences in anomalies according to normal ones by dimension reduction. However, training a single low-dimension latent space is limited to present the low-dimensional features due to the fact that the noise and irreverent features are mapped into this space, resulting in that the manifolds are not discriminative for detecting anomalies. To address this problem, a new autoencoder framework is proposed in this study with two trainable mutually orthogonal complementary subspaces in the latent space, by latent subspace projection (LSP) mechanism, which is named as LSP-CAE. Specifically, latent subspace projection is used to train the latent image subspace (LIS) and the latent kernel subspace (LKS) in the latent space of the autoencoder-like model respectively, which can enhance learning power of different features from the input instance. The features of normal data are projected into the latent image subspace, while the latent kernel subspace is trained to extract the irrelevant information according to normal features by end-to-end training. To verify the generality and effectiveness of the proposed method, we replace the convolutional network with the fully-connected network contucted in the real-world medical datasets. The anomaly score based on projection norms in two subspace is used to evaluate the anomalies in the testing. Consequently, our proposed method can achieve the best performance according to four public datasets in comparison of the state-of-the-art methods.
Image anomaly detection (AD) is widely researched on many occasions in computer vision tasks. High-dimensional data, such as image data, with noise and complex background is still challenging to detect anomalies under the situation that imbalanced or incomplete data are available. Some deep learning methods can be trained in an unsupervised way and map the original input into low-dimensional manifolds to predict larger differences in anomalies according to normal ones by dimension reduction. However, training a single low-dimension latent space is limited to present the low-dimensional features due to the fact that the noise and irreverent features are mapped into this space, resulting in that the manifolds are not discriminative for detecting anomalies. To address this problem, a new autoencoder framework is proposed in this study with two trainable mutually orthogonal complementary subspaces in the latent space, by latent subspace projection (LSP) mechanism, which is named as LSP-CAE. Specifically, latent subspace projection is used to train the latent image subspace (LIS) and the latent kernel subspace (LKS) in the latent space of the autoencoder-like model respectively, which can enhance learning power of different features from the input instance. The features of normal data are projected into the latent image subspace, while the latent kernel subspace is trained to extract the irrelevant information according to normal features by end-to-end training. To verify the generality and effectiveness of the proposed method, we replace the convolutional network with the fully-connected network contucted in the real-world medical datasets. The anomaly score based on projection norms in two subspace is used to evaluate the anomalies in the testing. Consequently, our proposed method can achieve the best performance according to four public datasets in comparison of the state-of-the-art methods. •We propose a latent subspace projection (LSP) constraints by incorporating deep encoder-decoder structure.•Complementary orthogonal subspaces can be trained in the end-to-end fashion to learn discriminative latent manifolds.•The proposed method shows the best performance in comparison of the state-of-the-art methods on public datasets.
Author Li, Chen
Yu, Qien
Zhu, Ye
Kurita, Takio
Author_xml – sequence: 1
  givenname: Qien
  surname: Yu
  fullname: Yu, Qien
  email: qienyu33@gmail.com
  organization: School of Information Science and Engineering, Chongqing Jiaotong University, Chongqing, 400074, China
– sequence: 2
  givenname: Chen
  surname: Li
  fullname: Li, Chen
  email: lichen7283@gmail.com
  organization: Graduate School of Science, Nagoya University, Chikusa, Nagoya, 464-8602, Japan
– sequence: 3
  givenname: Ye
  surname: Zhu
  fullname: Zhu, Ye
  email: ye.zhu@ieee.org
  organization: School of Information Technology, Deakin University, Victoria 3125, Australia
– sequence: 4
  givenname: Takio
  surname: Kurita
  fullname: Kurita, Takio
  email: tkurita@hiroshima-u.ac.jp
  organization: Graduate School of Advanced Science and Engineering, Hiroshima University, Higashi-hiroshima, Hiroshima, 739-8521, Japan
BackLink https://www.ncbi.nlm.nih.gov/pubmed/37120080$$D View this record in MEDLINE/PubMed
BookMark eNqFkU2L1TAUhoOMOB_6CwTJ0k3rSZrbpgsXclFHGHChrkM-TjGXNrkm6cD995N6RxcunFUOyfscwvNek4sQAxLymkHLgPXvDu1pwfKz5cC7FkQLMDwjVwzGXTOyDi62WfTN9nxJrnM-AADjg3xBLruBcQAJV-TbPob7OK_Fx6BnqtcSMdjoMFGjMzoaA511wVBoXk0-aov0mOIB7UbQKSaqQ1z0fKIOy_n2JXk-6Tnjq8fzhvz49PH7_ra5-_r5y_7DXWPFTpZG6x7lDtCZSVrnzICGIQNj9IRoQHAcJWf9xDgTXPe858YygZ22DDvpeHdD3p731g_9WjEXtfhscZ51wLhm1bGd4D0bpXgyyiUMI4xVWI2-eYyuZkGnjskvOp3UH2c10J0DNsWcE05_IwzU1ow6qN_NqE29AqFqM5Ua_6GsL3rTVZL28xPs-zOL1ea9x6Sy9bUmdD5V5cpF_1_-AYJQrBw
CitedBy_id crossref_primary_10_1016_j_patrec_2023_10_001
crossref_primary_10_1016_j_ymeth_2024_04_002
crossref_primary_10_1016_j_ymeth_2023_06_003
Cites_doi 10.1109/ACCESS.2018.2848210
10.1016/j.knosys.2019.07.008
10.1016/j.neucom.2021.04.033
10.1016/j.knosys.2018.11.030
10.1109/TNNLS.2014.2329534
10.1016/j.cviu.2018.02.006
10.1109/TIE.1930.896476
10.1016/j.neucom.2021.04.089
10.1587/transinf.2018OFP0007
10.1109/TSMCC.2012.2215319
10.1016/j.cviu.2011.03.003
10.1016/j.neucom.2019.08.044
10.1109/TIP.2019.2917862
10.1049/ip-vis:19990428
10.1109/TITS.2007.903444
10.1016/j.knosys.2019.105187
10.1016/j.knosys.2019.03.001
ContentType Journal Article
Copyright 2023 Elsevier Inc.
Copyright © 2023 Elsevier Inc. All rights reserved.
Copyright_xml – notice: 2023 Elsevier Inc.
– notice: Copyright © 2023 Elsevier Inc. All rights reserved.
DBID AAYXX
CITATION
CGR
CUY
CVF
ECM
EIF
NPM
7X8
7S9
L.6
DOI 10.1016/j.ymeth.2023.04.007
DatabaseName CrossRef
Medline
MEDLINE
MEDLINE (Ovid)
MEDLINE
MEDLINE
PubMed
MEDLINE - Academic
AGRICOLA
AGRICOLA - Academic
DatabaseTitle CrossRef
MEDLINE
Medline Complete
MEDLINE with Full Text
PubMed
MEDLINE (Ovid)
MEDLINE - Academic
AGRICOLA
AGRICOLA - Academic
DatabaseTitleList MEDLINE
AGRICOLA
MEDLINE - Academic

Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod fulltext_linktorsrc
Discipline Anatomy & Physiology
Chemistry
EISSN 1095-9130
EndPage 59
ExternalDocumentID 37120080
10_1016_j_ymeth_2023_04_007
S1046202323000671
Genre Journal Article
GroupedDBID ---
--K
--M
-~X
.GJ
.~1
0R~
123
1B1
1RT
1~.
1~5
29M
4.4
457
4G.
53G
5VS
7-5
71M
8P~
9JM
AABNK
AACTN
AAEDT
AAEDW
AAIAV
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAQXK
AAXUO
ABEFU
ABFNM
ABFRF
ABGSF
ABJNI
ABMAC
ABUDA
ABXDB
ABYKQ
ACDAQ
ACGFO
ACGFS
ACRLP
ADBBV
ADEZE
ADFGL
ADMUD
ADUVX
AEBSH
AEFWE
AEHWI
AEKER
AENEX
AFKWA
AFTJW
AFXIZ
AGHFR
AGRDE
AGUBO
AGYEJ
AHHHB
AIEXJ
AIKHN
AITUG
AJBFU
AJOXV
ALMA_UNASSIGNED_HOLDINGS
AMFUW
AMRAJ
ASPBG
AVWKF
AXJTR
AZFZN
BKOJK
BLXMC
CAG
COF
CS3
DM4
DOVZS
DU5
EBS
EFBJH
EFLBG
EJD
EO8
EO9
EP2
EP3
F5P
FDB
FEDTE
FGOYB
FIRID
FNPLU
FYGXN
G-2
G-Q
GBLVA
HLW
HMG
HVGLF
HZ~
IHE
J1W
K-O
KOM
LG5
LX2
LZ5
M41
MO0
N9A
O-L
O9-
OAUVE
OZT
P-8
P-9
P2P
PC.
Q38
R2-
RIG
ROL
RPZ
SBG
SCC
SDF
SDG
SDP
SES
SEW
SIN
SPCBC
SSU
SSZ
T5K
WUQ
XPP
Y6R
ZGI
ZMT
ZU3
~G-
9DU
AAHBH
AATTM
AAXKI
AAYWO
AAYXX
ABWVN
ACLOT
ACRPL
ACVFH
ADCNI
ADNMO
AEIPS
AEUPX
AFJKZ
AFPUW
AGQPQ
AIGII
AIIUN
AKBMS
AKRWK
AKYEP
ANKPU
APXCP
CITATION
EFKBS
~HD
CGR
CUY
CVF
ECM
EIF
NPM
PKN
7X8
7S9
L.6
ID FETCH-LOGICAL-c458t-aa6e850edbf8cddb7eb1e10bbafeeb042e98216f12142a6262bc14e3ac1e38d23
ISICitedReferencesCount 5
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001002260400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1046-2023
1095-9130
IngestDate Sun Sep 28 11:36:19 EDT 2025
Thu Oct 02 14:35:48 EDT 2025
Wed Feb 19 02:23:58 EST 2025
Sat Nov 29 07:07:27 EST 2025
Tue Nov 18 21:29:05 EST 2025
Fri Feb 23 02:37:11 EST 2024
IsPeerReviewed true
IsScholarly true
Keywords Subspace detection
Autoencoder
Latent subspace projection
Anomaly detection
Language English
License Copyright © 2023 Elsevier Inc. All rights reserved.
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c458t-aa6e850edbf8cddb7eb1e10bbafeeb042e98216f12142a6262bc14e3ac1e38d23
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
PMID 37120080
PQID 2807909104
PQPubID 23479
PageCount 12
ParticipantIDs proquest_miscellaneous_3154261984
proquest_miscellaneous_2807909104
pubmed_primary_37120080
crossref_primary_10_1016_j_ymeth_2023_04_007
crossref_citationtrail_10_1016_j_ymeth_2023_04_007
elsevier_sciencedirect_doi_10_1016_j_ymeth_2023_04_007
PublicationCentury 2000
PublicationDate June 2023
2023-06-00
20230601
PublicationDateYYYYMMDD 2023-06-01
PublicationDate_xml – month: 06
  year: 2023
  text: June 2023
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle Methods (San Diego, Calif.)
PublicationTitleAlternate Methods
PublicationYear 2023
Publisher Elsevier Inc
Publisher_xml – name: Elsevier Inc
References Russo, Disch, Blumensaat, Villez (br0340)
Zhou, Yang, Fujita, Chen, Wen (br0130) 2020; 187
Sabokrou, Fayyaz, Fathy, Moayed, Klette (br0260) 2018; 172
Reynolds (br0490) 2009
Pidhorskyi, Almohsen, Doretto (br0250) 2018; 31
Zhang, Bi, Xu, Ramentol, Fan, Qiao, Fujita (br0120) 2019; 174
Ruff, Vandermeulen, Goernitz, Deecke, Siddiqui, Binder, Müller, Kloft (br0150) 2018
Wang, Du, Lin, Cui, Shen, Yang (br0170) 2020; 190
Yu, Kavitha, Kurita (br0290) 2021; 450
Dufrenois (br0110) 2014; 26
Sodemann, Ross, Borghetti (br0060) 2012; 42
Roberts (br0100) 1999; 146
Beggel, Pfeiffer, Bischl (br0270) 2019
Schreyer, Sattarov, Schulze, Reimer, Borth (br0370)
Otomo, Kobayashi, Fukuda, Esaki (br0240) 2019; 102
Li, Chang (br0350) 2019; 369
Oza, Patel (br0040) 2019
Perera, Nallapati, Xiang (br0160) 2019
Chen, Yeo, Lee, Lau (br0200) 2018
Perera, Patel (br0050) 2018
Fan, Zhang, Wang, Xi, Li (br0280) 2020
Wang, Wong, Miner (br0320) 2004
Ramaswamy, Rastogi, Shim (br0330) 2000
Hasan, Choi, Neumann, Roy-Chowdhury, Davis (br0080) 2016
Gandhi, Trivedi (br0020) 2007; 8
Kumar (br0010) 2008; 55
Goodfellow, Pouget-Abadie, Mirza, Xu, Warde-Farley, Ozair, Courville, Bengio (br0360) 2014; 27
Akçay, Atapour-Abarghouei, Breckon (br0420) 2019
D. Dua, C. Graff, et al., Uci machine learning repository.
Gautam, Balaji, Sudharsan, Tiwari, Ahuja (br0180) 2019; 165
Kumagai, Iwata, Fujiwara (br0230) 2019; 32
Lemaître, Nogueira, Aridas (br0140) 2017; 18
Abati, Porrello, Calderara, Cucchiara (br0380) 2019
Van den Oord, Kalchbrenner, Espeholt, Vinyals, Graves (br0430) 2016; 29
Xiao, Rasul, Vollgraf (br0390)
Calderara, Heinemann, Prati, Cucchiara, Tishby (br0070) 2011; 115
Zhou, Liang, Zhang, Zhang, Song (br0190) 2021; 453
Latecki, Lazarevic, Pokrajac (br0310) 2007
Schlegl, Seeböck, Waldstein, Schmidt-Erfurth, Langs (br0090) 2017
Abe, Zadrozny, Langford (br0460) 2006
Sun, Wang, Xiong, Shao (br0220) 2018; 6
Kingma, Welling (br0210)
Krizhevsky, Hinton (br0400) 2009
Coates, Ng, Lee (br0410) 2011
Perera, Patel (br0030) 2019; 28
Kingma, Ba (br0440)
Li, Zhao, Botta, Ionescu, Hu (br0480) 2020
Dufrenois (10.1016/j.ymeth.2023.04.007_br0110) 2014; 26
Otomo (10.1016/j.ymeth.2023.04.007_br0240) 2019; 102
Oza (10.1016/j.ymeth.2023.04.007_br0040) 2019
Wang (10.1016/j.ymeth.2023.04.007_br0170) 2020; 190
Sun (10.1016/j.ymeth.2023.04.007_br0220) 2018; 6
Calderara (10.1016/j.ymeth.2023.04.007_br0070) 2011; 115
Zhang (10.1016/j.ymeth.2023.04.007_br0120) 2019; 174
Li (10.1016/j.ymeth.2023.04.007_br0480) 2020
Kingma (10.1016/j.ymeth.2023.04.007_br0440)
Kumagai (10.1016/j.ymeth.2023.04.007_br0230) 2019; 32
Goodfellow (10.1016/j.ymeth.2023.04.007_br0360) 2014; 27
Kingma (10.1016/j.ymeth.2023.04.007_br0210)
Perera (10.1016/j.ymeth.2023.04.007_br0030) 2019; 28
Chen (10.1016/j.ymeth.2023.04.007_br0200) 2018
Perera (10.1016/j.ymeth.2023.04.007_br0050) 2018
Lemaître (10.1016/j.ymeth.2023.04.007_br0140) 2017; 18
Coates (10.1016/j.ymeth.2023.04.007_br0410) 2011
Kumar (10.1016/j.ymeth.2023.04.007_br0010) 2008; 55
Pidhorskyi (10.1016/j.ymeth.2023.04.007_br0250) 2018; 31
Perera (10.1016/j.ymeth.2023.04.007_br0160) 2019
Xiao (10.1016/j.ymeth.2023.04.007_br0390)
Krizhevsky (10.1016/j.ymeth.2023.04.007_br0400) 2009
Fan (10.1016/j.ymeth.2023.04.007_br0280) 2020
Schlegl (10.1016/j.ymeth.2023.04.007_br0090) 2017
Van den Oord (10.1016/j.ymeth.2023.04.007_br0430) 2016; 29
10.1016/j.ymeth.2023.04.007_br0450
Wang (10.1016/j.ymeth.2023.04.007_br0320) 2004
Zhou (10.1016/j.ymeth.2023.04.007_br0130) 2020; 187
Zhou (10.1016/j.ymeth.2023.04.007_br0190) 2021; 453
Li (10.1016/j.ymeth.2023.04.007_br0350) 2019; 369
Akçay (10.1016/j.ymeth.2023.04.007_br0420) 2019
Beggel (10.1016/j.ymeth.2023.04.007_br0270) 2019
Schreyer (10.1016/j.ymeth.2023.04.007_br0370)
Abati (10.1016/j.ymeth.2023.04.007_br0380) 2019
Hasan (10.1016/j.ymeth.2023.04.007_br0080) 2016
Gautam (10.1016/j.ymeth.2023.04.007_br0180) 2019; 165
Ramaswamy (10.1016/j.ymeth.2023.04.007_br0330) 2000
Sabokrou (10.1016/j.ymeth.2023.04.007_br0260) 2018; 172
Reynolds (10.1016/j.ymeth.2023.04.007_br0490) 2009
Roberts (10.1016/j.ymeth.2023.04.007_br0100) 1999; 146
Ruff (10.1016/j.ymeth.2023.04.007_br0150) 2018
Sodemann (10.1016/j.ymeth.2023.04.007_br0060) 2012; 42
Abe (10.1016/j.ymeth.2023.04.007_br0460) 2006
Russo (10.1016/j.ymeth.2023.04.007_br0340)
Gandhi (10.1016/j.ymeth.2023.04.007_br0020) 2007; 8
Yu (10.1016/j.ymeth.2023.04.007_br0290) 2021; 450
Latecki (10.1016/j.ymeth.2023.04.007_br0310) 2007
References_xml – volume: 32
  year: 2019
  ident: br0230
  article-title: Transfer anomaly detection by inferring latent domain representations
  publication-title: Adv. Neural Inf. Process. Syst.
– start-page: 61
  year: 2007
  end-page: 75
  ident: br0310
  article-title: Outlier detection with kernel density functions
  publication-title: International Workshop on Machine Learning and Data Mining in Pattern Recognition
– start-page: 427
  year: 2000
  end-page: 438
  ident: br0330
  article-title: Efficient algorithms for mining outliers from large data sets
  publication-title: Proceedings of the 2000 ACM SIGMOD International Conference on Management of Data
– volume: 27
  year: 2014
  ident: br0360
  article-title: Generative adversarial nets
  publication-title: Adv. Neural Inf. Process. Syst.
– start-page: 733
  year: 2016
  end-page: 742
  ident: br0080
  article-title: Learning temporal regularity in video sequences
  publication-title: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition
– volume: 165
  start-page: 241
  year: 2019
  end-page: 252
  ident: br0180
  article-title: Localized multiple kernel learning for anomaly detection: one-class classification
  publication-title: Knowl.-Based Syst.
– volume: 450
  start-page: 372
  year: 2021
  end-page: 388
  ident: br0290
  article-title: Autoencoder framework based on orthogonal projection constraints improves anomalies detection
  publication-title: Neurocomputing
– start-page: 215
  year: 2011
  end-page: 223
  ident: br0410
  article-title: An analysis of single-layer networks in unsupervised feature learning
  publication-title: Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, JMLR Workshop and Conference Proceedings
– volume: 190
  year: 2020
  ident: br0170
  article-title: Advae: a self-adversarial variational autoencoder with Gaussian anomaly prior knowledge for anomaly detection
  publication-title: Knowl.-Based Syst.
– volume: 29
  year: 2016
  ident: br0430
  article-title: Conditional image generation with pixelcnn decoders
  publication-title: Adv. Neural Inf. Process. Syst.
– start-page: 206
  year: 2019
  end-page: 222
  ident: br0270
  article-title: Robust anomaly detection in images using adversarial autoencoders
  publication-title: Joint European Conference on Machine Learning and Knowledge Discovery in Databases
– start-page: 1
  year: 2019
  end-page: 8
  ident: br0420
  article-title: Skip-ganomaly: skip connected and adversarially trained encoder-decoder anomaly detection
  publication-title: 2019 International Joint Conference on Neural Networks (IJCNN)
– volume: 369
  start-page: 92
  year: 2019
  end-page: 105
  ident: br0350
  article-title: Video anomaly detection and localization via multivariate Gaussian fully convolution adversarial autoencoder
  publication-title: Neurocomputing
– start-page: 1
  year: 2018
  end-page: 5
  ident: br0200
  article-title: Autoencoder-based network anomaly detection
  publication-title: 2018 Wireless Telecommunications Symposium (WTS)
– year: 2009
  ident: br0400
  article-title: Learning Multiple Layers of Features from Tiny Images
– volume: 102
  start-page: 1644
  year: 2019
  end-page: 1652
  ident: br0240
  article-title: Latent variable based anomaly detection in network system logs
  publication-title: IEICE Trans. Inf. Syst.
– volume: 31
  year: 2018
  ident: br0250
  article-title: Generative probabilistic novelty detection with adversarial autoencoders
  publication-title: Adv. Neural Inf. Process. Syst.
– start-page: 659
  year: 2009
  end-page: 663
  ident: br0490
  article-title: Gaussian mixture models
  publication-title: Encyclopedia of Biometrics, vol. 741
– start-page: 2898
  year: 2019
  end-page: 2906
  ident: br0160
  article-title: Ocgan: one-class novelty detection using gans with constrained latent representations
  publication-title: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition
– start-page: 358
  year: 2004
  end-page: 364
  ident: br0320
  article-title: Anomaly intrusion detection using one class svm
  publication-title: Proceedings from the Fifth Annual IEEE SMC Information Assurance Workshop, 2004
– volume: 146
  start-page: 124
  year: 1999
  end-page: 129
  ident: br0100
  article-title: Novelty detection using extreme value statistics
  publication-title: IEE Proc., Vis. Image Signal Process.
– ident: br0440
  article-title: Adam: a method for stochastic optimization
– volume: 55
  start-page: 348
  year: 2008
  end-page: 363
  ident: br0010
  article-title: Computer-vision-based fabric defect detection: a survey
  publication-title: IEEE Trans. Ind. Electron.
– volume: 28
  start-page: 5450
  year: 2019
  end-page: 5463
  ident: br0030
  article-title: Learning deep features for one-class classification
  publication-title: IEEE Trans. Image Process.
– volume: 115
  start-page: 1099
  year: 2011
  end-page: 1111
  ident: br0070
  article-title: Detecting anomalies in people's trajectories using spectral graph analysis
  publication-title: Comput. Vis. Image Underst.
– volume: 453
  start-page: 131
  year: 2021
  end-page: 140
  ident: br0190
  article-title: Vae-based deep svdd for anomaly detection
  publication-title: Neurocomputing
– volume: 42
  start-page: 1257
  year: 2012
  end-page: 1272
  ident: br0060
  article-title: A review of anomaly detection in automated surveillance
  publication-title: IEEE Trans. Syst. Man Cybern., Part C, Appl. Rev.
– ident: br0390
  article-title: Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms
– start-page: 146
  year: 2017
  end-page: 157
  ident: br0090
  article-title: Unsupervised anomaly detection with generative adversarial networks to guide marker discovery
  publication-title: International Conference on Information Processing in Medical Imaging
– volume: 172
  start-page: 88
  year: 2018
  end-page: 97
  ident: br0260
  article-title: Deep-anomaly: fully convolutional neural network for fast anomaly detection in crowded scenes
  publication-title: Comput. Vis. Image Underst.
– start-page: 1118
  year: 2020
  end-page: 1123
  ident: br0480
  article-title: Copod: copula-based outlier detection
  publication-title: 2020 IEEE International Conference on Data Mining (ICDM)
– volume: 174
  start-page: 137
  year: 2019
  end-page: 143
  ident: br0120
  article-title: Multi-imbalance: an open-source software for multi-class imbalance learning
  publication-title: Knowl.-Based Syst.
– start-page: 1
  year: 2019
  end-page: 8
  ident: br0040
  article-title: Active authentication using an autoencoder regularized cnn-based one-class classifier
  publication-title: 2019 14th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2019)
– volume: 26
  start-page: 982
  year: 2014
  end-page: 994
  ident: br0110
  article-title: A one-class kernel Fisher criterion for outlier detection
  publication-title: IEEE Trans. Neural Netw. Learn. Syst.
– ident: br0370
  article-title: Detection of accounting anomalies in the latent space using adversarial autoencoder neural networks
– volume: 8
  start-page: 413
  year: 2007
  end-page: 430
  ident: br0020
  article-title: Pedestrian protection systems: issues, survey, and challenges
  publication-title: IEEE Trans. Intell. Transp. Syst.
– start-page: 4393
  year: 2018
  end-page: 4402
  ident: br0150
  article-title: Deep one-class classification
  publication-title: International Conference on Machine Learning, PMLR
– ident: br0340
  article-title: Anomaly detection using deep autoencoders for in-situ wastewater systems monitoring data
– ident: br0210
  article-title: Auto-encoding variational Bayes
– reference: D. Dua, C. Graff, et al., Uci machine learning repository.
– volume: 187
  year: 2020
  ident: br0130
  article-title: Deep learning fault diagnosis method based on global optimization gan for unbalanced data
  publication-title: Knowl.-Based Syst.
– start-page: 504
  year: 2006
  end-page: 509
  ident: br0460
  article-title: Outlier detection by active learning
  publication-title: Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
– start-page: 1
  year: 2018
  end-page: 8
  ident: br0050
  article-title: Dual-minimax probability machines for one-class mobile active authentication
  publication-title: 2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems (BTAS)
– start-page: 688
  year: 2020
  end-page: 700
  ident: br0280
  article-title: Correlation-aware deep generative model for unsupervised anomaly detection
  publication-title: Pacific-Asia Conference on Knowledge Discovery and Data Mining
– volume: 6
  start-page: 33353
  year: 2018
  end-page: 33361
  ident: br0220
  article-title: Learning sparse representation with variational auto-encoder for anomaly detection
  publication-title: IEEE Access
– volume: 18
  start-page: 559
  year: 2017
  end-page: 563
  ident: br0140
  article-title: Imbalanced-learn: a python toolbox to tackle the curse of imbalanced datasets in machine learning
  publication-title: J. Mach. Learn. Res.
– start-page: 481
  year: 2019
  end-page: 490
  ident: br0380
  article-title: Latent space autoregression for novelty detection
  publication-title: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition
– volume: 6
  start-page: 33353
  year: 2018
  ident: 10.1016/j.ymeth.2023.04.007_br0220
  article-title: Learning sparse representation with variational auto-encoder for anomaly detection
  publication-title: IEEE Access
  doi: 10.1109/ACCESS.2018.2848210
– volume: 187
  year: 2020
  ident: 10.1016/j.ymeth.2023.04.007_br0130
  article-title: Deep learning fault diagnosis method based on global optimization gan for unbalanced data
  publication-title: Knowl.-Based Syst.
  doi: 10.1016/j.knosys.2019.07.008
– ident: 10.1016/j.ymeth.2023.04.007_br0370
– ident: 10.1016/j.ymeth.2023.04.007_br0210
– volume: 450
  start-page: 372
  year: 2021
  ident: 10.1016/j.ymeth.2023.04.007_br0290
  article-title: Autoencoder framework based on orthogonal projection constraints improves anomalies detection
  publication-title: Neurocomputing
  doi: 10.1016/j.neucom.2021.04.033
– volume: 32
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0230
  article-title: Transfer anomaly detection by inferring latent domain representations
  publication-title: Adv. Neural Inf. Process. Syst.
– ident: 10.1016/j.ymeth.2023.04.007_br0450
– year: 2009
  ident: 10.1016/j.ymeth.2023.04.007_br0400
– volume: 165
  start-page: 241
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0180
  article-title: Localized multiple kernel learning for anomaly detection: one-class classification
  publication-title: Knowl.-Based Syst.
  doi: 10.1016/j.knosys.2018.11.030
– start-page: 659
  year: 2009
  ident: 10.1016/j.ymeth.2023.04.007_br0490
  article-title: Gaussian mixture models
– start-page: 146
  year: 2017
  ident: 10.1016/j.ymeth.2023.04.007_br0090
  article-title: Unsupervised anomaly detection with generative adversarial networks to guide marker discovery
– ident: 10.1016/j.ymeth.2023.04.007_br0390
– start-page: 1
  year: 2018
  ident: 10.1016/j.ymeth.2023.04.007_br0050
  article-title: Dual-minimax probability machines for one-class mobile active authentication
– volume: 26
  start-page: 982
  issue: 5
  year: 2014
  ident: 10.1016/j.ymeth.2023.04.007_br0110
  article-title: A one-class kernel Fisher criterion for outlier detection
  publication-title: IEEE Trans. Neural Netw. Learn. Syst.
  doi: 10.1109/TNNLS.2014.2329534
– volume: 172
  start-page: 88
  year: 2018
  ident: 10.1016/j.ymeth.2023.04.007_br0260
  article-title: Deep-anomaly: fully convolutional neural network for fast anomaly detection in crowded scenes
  publication-title: Comput. Vis. Image Underst.
  doi: 10.1016/j.cviu.2018.02.006
– volume: 55
  start-page: 348
  issue: 1
  year: 2008
  ident: 10.1016/j.ymeth.2023.04.007_br0010
  article-title: Computer-vision-based fabric defect detection: a survey
  publication-title: IEEE Trans. Ind. Electron.
  doi: 10.1109/TIE.1930.896476
– volume: 453
  start-page: 131
  year: 2021
  ident: 10.1016/j.ymeth.2023.04.007_br0190
  article-title: Vae-based deep svdd for anomaly detection
  publication-title: Neurocomputing
  doi: 10.1016/j.neucom.2021.04.089
– volume: 102
  start-page: 1644
  issue: 9
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0240
  article-title: Latent variable based anomaly detection in network system logs
  publication-title: IEICE Trans. Inf. Syst.
  doi: 10.1587/transinf.2018OFP0007
– start-page: 427
  year: 2000
  ident: 10.1016/j.ymeth.2023.04.007_br0330
  article-title: Efficient algorithms for mining outliers from large data sets
– start-page: 61
  year: 2007
  ident: 10.1016/j.ymeth.2023.04.007_br0310
  article-title: Outlier detection with kernel density functions
– start-page: 215
  year: 2011
  ident: 10.1016/j.ymeth.2023.04.007_br0410
  article-title: An analysis of single-layer networks in unsupervised feature learning
– volume: 18
  start-page: 559
  issue: 1
  year: 2017
  ident: 10.1016/j.ymeth.2023.04.007_br0140
  article-title: Imbalanced-learn: a python toolbox to tackle the curse of imbalanced datasets in machine learning
  publication-title: J. Mach. Learn. Res.
– volume: 27
  year: 2014
  ident: 10.1016/j.ymeth.2023.04.007_br0360
  article-title: Generative adversarial nets
  publication-title: Adv. Neural Inf. Process. Syst.
– ident: 10.1016/j.ymeth.2023.04.007_br0440
– start-page: 4393
  year: 2018
  ident: 10.1016/j.ymeth.2023.04.007_br0150
  article-title: Deep one-class classification
– start-page: 1
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0040
  article-title: Active authentication using an autoencoder regularized cnn-based one-class classifier
– start-page: 206
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0270
  article-title: Robust anomaly detection in images using adversarial autoencoders
– volume: 42
  start-page: 1257
  issue: 6
  year: 2012
  ident: 10.1016/j.ymeth.2023.04.007_br0060
  article-title: A review of anomaly detection in automated surveillance
  publication-title: IEEE Trans. Syst. Man Cybern., Part C, Appl. Rev.
  doi: 10.1109/TSMCC.2012.2215319
– volume: 115
  start-page: 1099
  issue: 8
  year: 2011
  ident: 10.1016/j.ymeth.2023.04.007_br0070
  article-title: Detecting anomalies in people's trajectories using spectral graph analysis
  publication-title: Comput. Vis. Image Underst.
  doi: 10.1016/j.cviu.2011.03.003
– start-page: 688
  year: 2020
  ident: 10.1016/j.ymeth.2023.04.007_br0280
  article-title: Correlation-aware deep generative model for unsupervised anomaly detection
– start-page: 2898
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0160
  article-title: Ocgan: one-class novelty detection using gans with constrained latent representations
– start-page: 1
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0420
  article-title: Skip-ganomaly: skip connected and adversarially trained encoder-decoder anomaly detection
– start-page: 481
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0380
  article-title: Latent space autoregression for novelty detection
– start-page: 733
  year: 2016
  ident: 10.1016/j.ymeth.2023.04.007_br0080
  article-title: Learning temporal regularity in video sequences
– volume: 369
  start-page: 92
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0350
  article-title: Video anomaly detection and localization via multivariate Gaussian fully convolution adversarial autoencoder
  publication-title: Neurocomputing
  doi: 10.1016/j.neucom.2019.08.044
– start-page: 504
  year: 2006
  ident: 10.1016/j.ymeth.2023.04.007_br0460
  article-title: Outlier detection by active learning
– start-page: 1
  year: 2018
  ident: 10.1016/j.ymeth.2023.04.007_br0200
  article-title: Autoencoder-based network anomaly detection
– volume: 28
  start-page: 5450
  issue: 11
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0030
  article-title: Learning deep features for one-class classification
  publication-title: IEEE Trans. Image Process.
  doi: 10.1109/TIP.2019.2917862
– start-page: 1118
  year: 2020
  ident: 10.1016/j.ymeth.2023.04.007_br0480
  article-title: Copod: copula-based outlier detection
– start-page: 358
  year: 2004
  ident: 10.1016/j.ymeth.2023.04.007_br0320
  article-title: Anomaly intrusion detection using one class svm
– volume: 29
  year: 2016
  ident: 10.1016/j.ymeth.2023.04.007_br0430
  article-title: Conditional image generation with pixelcnn decoders
  publication-title: Adv. Neural Inf. Process. Syst.
– volume: 146
  start-page: 124
  issue: 3
  year: 1999
  ident: 10.1016/j.ymeth.2023.04.007_br0100
  article-title: Novelty detection using extreme value statistics
  publication-title: IEE Proc., Vis. Image Signal Process.
  doi: 10.1049/ip-vis:19990428
– volume: 31
  year: 2018
  ident: 10.1016/j.ymeth.2023.04.007_br0250
  article-title: Generative probabilistic novelty detection with adversarial autoencoders
  publication-title: Adv. Neural Inf. Process. Syst.
– volume: 8
  start-page: 413
  issue: 3
  year: 2007
  ident: 10.1016/j.ymeth.2023.04.007_br0020
  article-title: Pedestrian protection systems: issues, survey, and challenges
  publication-title: IEEE Trans. Intell. Transp. Syst.
  doi: 10.1109/TITS.2007.903444
– volume: 190
  year: 2020
  ident: 10.1016/j.ymeth.2023.04.007_br0170
  article-title: Advae: a self-adversarial variational autoencoder with Gaussian anomaly prior knowledge for anomaly detection
  publication-title: Knowl.-Based Syst.
  doi: 10.1016/j.knosys.2019.105187
– volume: 174
  start-page: 137
  year: 2019
  ident: 10.1016/j.ymeth.2023.04.007_br0120
  article-title: Multi-imbalance: an open-source software for multi-class imbalance learning
  publication-title: Knowl.-Based Syst.
  doi: 10.1016/j.knosys.2019.03.001
– ident: 10.1016/j.ymeth.2023.04.007_br0340
SSID ssj0001278
Score 2.459901
Snippet Image anomaly detection (AD) is widely researched on many occasions in computer vision tasks. High-dimensional data, such as image data, with noise and complex...
SourceID proquest
pubmed
crossref
elsevier
SourceType Aggregation Database
Index Database
Enrichment Source
Publisher
StartPage 48
SubjectTerms Algorithms
Anomaly detection
Autoencoder
computer vision
data collection
Latent subspace projection
Subspace detection
Title Convolutional autoencoder based on latent subspace projection for anomaly detection
URI https://dx.doi.org/10.1016/j.ymeth.2023.04.007
https://www.ncbi.nlm.nih.gov/pubmed/37120080
https://www.proquest.com/docview/2807909104
https://www.proquest.com/docview/3154261984
Volume 214
WOSCitedRecordID wos001002260400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVESC
  databaseName: Elsevier SD Freedom Collection Journals 2021
  customDbUrl:
  eissn: 1095-9130
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0001278
  issn: 1046-2023
  databaseCode: AIEXJ
  dateStart: 19950201
  isFulltext: true
  titleUrlDefault: https://www.sciencedirect.com
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3fb9MwELZYh2AvCDYG48dkJMTLSBQnaeI8VmUI0JhAK6jjJXISFzq1TrWm0_rfc2fHSdDUCR54iSLHcaJ8X8535_MdIa-Lgmc-TB2OCFgEBgoPnCTPCqeQUR6B-ir6umjf95P49JSPx8mXOn5-qcsJxErx6-tk8V-hhjYAG7fO_gPczaDQAOcAOhwBdjj-FfDDUl3VD8A8AKuqxFyVmDICZ6wCVwdmoGCq6mgJQgNMZtwqpd0xNupQqHIuZuujQlamtavBftYlp7Wv9gxkw7up_Gm8rbjJy-04Fs5X2Px12u41O5maBf625ccv3em8YReuKBl9dgSqbdn1SfhBGzvlSiNHPVMC0usKWp-FHVFpEmzekODGmXDhrrGCtouDuybBebc3fPHFXOMXxAwjOLx2OmuCDO2lLbLtx_2E98j24OPx-FMzUTM_5jYRlQ75u_HMHXLPjrJJb9lkl2j9ZPSQPKgNCzowhHhE7ki1S_YGSlTlfE3fUB3qq9dQdsn9oS3zt0fO_uAL7fCFar7QUlHDF2r5Qlu-UOALrflCG748Jt_eH4-GH5y60oaTh31eOUJEkvc9WWQTnheYcTtjknlZJiZSZiDXZcJ9Fk0YJugTYAP7Wc5CGYicyYAXfrBPeqpU8imhYRH3szyPZCRCrGQj_MQLhYh9HsmYSX5AfPsd07xOQ4_VUGapjTe8SDUOKeKQemEKOByQt81NC5OF5fbukQUorRVJoyCmwLLbb3xl4UwBBlw7E0qWq2WKSaMS1K3DzX0CMEfQIcGhzxPDheZtLY2ebbzynOy0v9IL0qsuV_IluZtfVdPl5SHZisf8sGbwbx7Lr3A
linkProvider Elsevier
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=Convolutional+autoencoder+based+on+latent+subspace+projection+for+anomaly+detection&rft.jtitle=Methods+%28San+Diego%2C+Calif.%29&rft.au=Yu%2C+Qien&rft.au=Li%2C+Chen&rft.au=Zhu%2C+Ye&rft.au=Kurita%2C+Takio&rft.date=2023-06-01&rft.eissn=1095-9130&rft.volume=214&rft.spage=48&rft_id=info:doi/10.1016%2Fj.ymeth.2023.04.007&rft_id=info%3Apmid%2F37120080&rft.externalDocID=37120080
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1046-2023&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1046-2023&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1046-2023&client=summon