Robust Spectral Anomaly Detection in EELS Spectral Images via 3D Convolutional Variational Autoencoders

A 3D Convolutional Variational Autoencoder (3D‐CVAE) is introduced for automated anomaly detection in electron energy‐loss spectroscopy spectrum imaging (EELS‐SI) data. This approach leverages the full 3D structure of EELS‐SI data to detect subtle spectral anomalies while preserving both spatial and...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Small (Weinheim an der Bergstrasse, Germany) Ročník 21; číslo 33; s. e2503019 - n/a
Hlavní autoři: Sultanov, Seyfal, Ayyubi, R. A. W., Buban, James P., Klie, Robert F.
Médium: Journal Article
Jazyk:angličtina
Vydáno: Germany Wiley Subscription Services, Inc 01.08.2025
John Wiley and Sons Inc
Témata:
ISSN:1613-6810, 1613-6829, 1613-6829
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract A 3D Convolutional Variational Autoencoder (3D‐CVAE) is introduced for automated anomaly detection in electron energy‐loss spectroscopy spectrum imaging (EELS‐SI) data. This approach leverages the full 3D structure of EELS‐SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross‐entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect‐free material. In exploring methods for anomaly detection, both the 3D‐CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L‐edge ΔE peak shifts designed to simulate material defects. These results show that 3D‐CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower‐dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise‐dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS‐SI data, particularly valuable for analyzing complex material systems. Automated anomaly detection is demonstrated for electron energy‐loss spectrum imaging in an atomic‐resolution scanning transmission electron microscope using an unsupervised learning approach. A 3D convolutional variational autoencoder is introduced and tested on the iron L‐edge spectra taken from a single‐crystal BiFeO3 sample. This approach is benchmarked against Principal Component Analysis and high reconstruction quality is demonstrated even in challenging, noise‐dominated spectral regions.
AbstractList A 3D Convolutional Variational Autoencoder (3D-CVAE) is introduced for automated anomaly detection in electron energy-loss spectroscopy spectrum imaging (EELS-SI) data. This approach leverages the full 3D structure of EELS-SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross-entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect-free material. In exploring methods for anomaly detection, both the 3D-CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L-edge ΔE peak shifts designed to simulate material defects. These results show that 3D-CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower-dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise-dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS-SI data, particularly valuable for analyzing complex material systems.
A 3D Convolutional Variational Autoencoder (3D-CVAE) is introduced for automated anomaly detection in electron energy-loss spectroscopy spectrum imaging (EELS-SI) data. This approach leverages the full 3D structure of EELS-SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross-entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect-free material. In exploring methods for anomaly detection, both the 3D-CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L-edge ΔE peak shifts designed to simulate material defects. These results show that 3D-CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower-dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise-dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS-SI data, particularly valuable for analyzing complex material systems.A 3D Convolutional Variational Autoencoder (3D-CVAE) is introduced for automated anomaly detection in electron energy-loss spectroscopy spectrum imaging (EELS-SI) data. This approach leverages the full 3D structure of EELS-SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross-entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect-free material. In exploring methods for anomaly detection, both the 3D-CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L-edge ΔE peak shifts designed to simulate material defects. These results show that 3D-CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower-dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise-dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS-SI data, particularly valuable for analyzing complex material systems.
A 3D Convolutional Variational Autoencoder (3D‐CVAE) is introduced for automated anomaly detection in electron energy‐loss spectroscopy spectrum imaging (EELS‐SI) data. This approach leverages the full 3D structure of EELS‐SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross‐entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect‐free material. In exploring methods for anomaly detection, both the 3D‐CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L ‐edge Δ E peak shifts designed to simulate material defects. These results show that 3D‐CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower‐dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise‐dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS‐SI data, particularly valuable for analyzing complex material systems.
A 3D Convolutional Variational Autoencoder (3D‐CVAE) is introduced for automated anomaly detection in electron energy‐loss spectroscopy spectrum imaging (EELS‐SI) data. This approach leverages the full 3D structure of EELS‐SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross‐entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect‐free material. In exploring methods for anomaly detection, both the 3D‐CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L‐edge ΔE peak shifts designed to simulate material defects. These results show that 3D‐CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower‐dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise‐dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS‐SI data, particularly valuable for analyzing complex material systems.
A 3D Convolutional Variational Autoencoder (3D‐CVAE) is introduced for automated anomaly detection in electron energy‐loss spectroscopy spectrum imaging (EELS‐SI) data. This approach leverages the full 3D structure of EELS‐SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross‐entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect‐free material. In exploring methods for anomaly detection, both the 3D‐CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L‐edge ΔE peak shifts designed to simulate material defects. These results show that 3D‐CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower‐dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise‐dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS‐SI data, particularly valuable for analyzing complex material systems. Automated anomaly detection is demonstrated for electron energy‐loss spectrum imaging in an atomic‐resolution scanning transmission electron microscope using an unsupervised learning approach. A 3D convolutional variational autoencoder is introduced and tested on the iron L‐edge spectra taken from a single‐crystal BiFeO3 sample. This approach is benchmarked against Principal Component Analysis and high reconstruction quality is demonstrated even in challenging, noise‐dominated spectral regions.
A 3D Convolutional Variational Autoencoder (3D‐CVAE) is introduced for automated anomaly detection in electron energy‐loss spectroscopy spectrum imaging (EELS‐SI) data. This approach leverages the full 3D structure of EELS‐SI data to detect subtle spectral anomalies while preserving both spatial and spectral correlations across the datacube. By employing cross‐entropy loss and training on bulk spectra, the model learns to reconstruct bulk features characteristic of the defect‐free material. In exploring methods for anomaly detection, both the 3D‐CVAE approach and principal component analysis (PCA) are evaluated, testing their performance using Fe L‐edge ΔE peak shifts designed to simulate material defects. These results show that 3D‐CVAE achieves superior anomaly detection and maintains consistent performance across various shift magnitudes. The method demonstrates clear bimodal separation between bulk and anomalous spectra, enabling reliable classification. Further analysis verifies that lower‐dimensional representations are robust to anomalies in the data. While performance advantages over PCA diminish with decreasing anomaly concentration, our method maintains high reconstruction quality even in challenging, noise‐dominated spectral regions. This approach provides a robust framework for unsupervised automated detection of spectral anomalies in EELS‐SI data, particularly valuable for analyzing complex material systems. Automated anomaly detection is demonstrated for electron energy‐loss spectrum imaging in an atomic‐resolution scanning transmission electron microscope using an unsupervised learning approach. A 3D convolutional variational autoencoder is introduced and tested on the iron L‐edge spectra taken from a single‐crystal BiFeO3 sample. This approach is benchmarked against Principal Component Analysis and high reconstruction quality is demonstrated even in challenging, noise‐dominated spectral regions.
Author Ayyubi, R. A. W.
Buban, James P.
Klie, Robert F.
Sultanov, Seyfal
AuthorAffiliation 1 Department of Computer Science University of Illinois Chicago Chicago IL 60607 USA
2 Department of Physics University of Illinois Chicago Chicago IL 60607 USA
AuthorAffiliation_xml – name: 2 Department of Physics University of Illinois Chicago Chicago IL 60607 USA
– name: 1 Department of Computer Science University of Illinois Chicago Chicago IL 60607 USA
Author_xml – sequence: 1
  givenname: Seyfal
  orcidid: 0009-0005-4948-7811
  surname: Sultanov
  fullname: Sultanov, Seyfal
  organization: University of Illinois Chicago
– sequence: 2
  givenname: R. A. W.
  surname: Ayyubi
  fullname: Ayyubi, R. A. W.
  organization: University of Illinois Chicago
– sequence: 3
  givenname: James P.
  surname: Buban
  fullname: Buban, James P.
  organization: University of Illinois Chicago
– sequence: 4
  givenname: Robert F.
  orcidid: 0000-0003-4773-6667
  surname: Klie
  fullname: Klie, Robert F.
  email: rfklie@uic.edu
  organization: University of Illinois Chicago
BackLink https://www.ncbi.nlm.nih.gov/pubmed/40619908$$D View this record in MEDLINE/PubMed
BookMark eNqFkUuP0zAUhS00iHnAliWKxGY2LX4lsVeo6hQYKQiJAlvLsW-LR45d7KSo_55ELR1gw8pXvt85OlfnGl2EGAChlwTPCcb0Te68n1NMS8wwkU_QFakIm1WCyovzTPAlus75AWNGKK-foUuOKyIlFldo-zm2Q-6L9Q5Mn7QvFiF22h-KO-jHHxdD4UKxWjXrR-S-01vIxd7pgt0Vyxj20Q8TOu6-6eT0aV4MfYRgooWUn6OnG-0zvDi9N-jru9WX5YdZ8-n9_XLRzAynlZzJasMZNaJloKHUltBWbKQouQDLKJW2ZhYEo9a0Jasl1ZU0uLRcspZYIQ27QW-Pvruh7cAaCFNktUuu0-mgonbq701w39U27hWhrKaclqPD7ckhxR8D5F51LhvwXgeIQ1ZjjLpiVckm9PU_6EMc0nj6RHEial5jMlKv_ox0zvK7hBGYHwGTYs4JNmeEYDW1rKaW1bnlUSCPgp_Ow-E_tFp_bJpH7S_qA6uv
Cites_doi 10.1103/PhysRevB.108.214428
10.1098/rspl.1895.0041
10.1088/2632-2153/acf6a9
10.1016/0304-3991(89)90304-5
10.1103/PhysRevLett.99.047203
10.1016/0921-4534(93)90501-G
10.1016/j.apsadv.2023.100523
10.1021/nl1022139
10.1038/416826a
10.1007/978-1-4419-7200-2
10.1177/14759217211073301
10.1023/A:1021428003587
10.1021/acs.chemmater.3c02773
10.1186/s40679-018-0052-y
10.1088/2632-2153/ad073b
10.1038/s41598-021-97668-8
10.3389/fnins.2021.652987
10.1109/ACCESS.2018.2848210
10.1126/science.1127647
10.1016/j.cviu.2020.102920
10.1126/science.aaa8415
10.1109/TSMC.1979.4310076
10.1038/nature14539
10.1103/RevModPhys.91.045002
10.1093/mam/ozae044.180
10.1109/5.726791
10.1109/TGRS.2019.2908756
10.1038/s42256-022-00555-8
10.1103/PhysRevLett.92.095502
10.1002/smll.202205977
10.1038/nature03644
10.1109/ACCESS.2021.3065838
10.1038/336565a0
10.1063/1.1330572
ContentType Journal Article
Copyright 2025 The Author(s). Small published by Wiley‐VCH GmbH
2025 The Author(s). Small published by Wiley‐VCH GmbH.
2025. This work is published under Creative Commons Attribution – Non-Commercial – No Derivatives License~http://creativecommons.org/licenses/by-nc-nd/3.0/ (the "License"). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Copyright_xml – notice: 2025 The Author(s). Small published by Wiley‐VCH GmbH
– notice: 2025 The Author(s). Small published by Wiley‐VCH GmbH.
– notice: 2025. This work is published under Creative Commons Attribution – Non-Commercial – No Derivatives License~http://creativecommons.org/licenses/by-nc-nd/3.0/ (the "License"). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
DBID 24P
AAYXX
CITATION
NPM
7SR
7U5
8BQ
8FD
JG9
L7M
7X8
5PM
DOI 10.1002/smll.202503019
DatabaseName Wiley Online Library Open Access
CrossRef
PubMed
Engineered Materials Abstracts
Solid State and Superconductivity Abstracts
METADEX
Technology Research Database
Materials Research Database
Advanced Technologies Database with Aerospace
MEDLINE - Academic
PubMed Central (Full Participant titles)
DatabaseTitle CrossRef
PubMed
Materials Research Database
Engineered Materials Abstracts
Solid State and Superconductivity Abstracts
Technology Research Database
Advanced Technologies Database with Aerospace
METADEX
MEDLINE - Academic
DatabaseTitleList PubMed
MEDLINE - Academic
CrossRef
Materials Research Database


Database_xml – sequence: 1
  dbid: 24P
  name: Wiley Online Library Open Access
  url: https://authorservices.wiley.com/open-science/open-access/browse-journals.html
  sourceTypes: Publisher
– sequence: 2
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 3
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 1613-6829
EndPage n/a
ExternalDocumentID PMC12372425
40619908
10_1002_smll_202503019
SMLL202503019
Genre researchArticle
Journal Article
GrantInformation_xml – fundername: Alliance for Sustainable Energy, LLC
– fundername: Solar Energy Technologies Program
  funderid: 37989
– fundername: U.S. Department of Energy's Office of Energy Efficiency and Renewable Energy (EERE)
  funderid: 37989
– fundername: Solar Energy Technologies Program
  grantid: 37989
– fundername: U.S. Department of Energy's Office of Energy Efficiency and Renewable Energy (EERE)
  grantid: 37989
– fundername: ;
– fundername: ;
  grantid: 37989
GroupedDBID ---
05W
0R~
123
1L6
1OC
24P
33P
3SF
3WU
4.4
50Y
52U
5VS
66C
8-0
8-1
8UM
AAESR
AAEVG
AAHQN
AAIHA
AAMMB
AAMNL
AANLZ
AAONW
AAXRX
AAYCA
AAZKR
ABCUV
ABIJN
ABJNI
ABLJU
ABRTZ
ACAHQ
ACCZN
ACFBH
ACGFS
ACIWK
ACPOU
ACXBN
ACXQS
ADBBV
ADEOM
ADIZJ
ADKYN
ADMGS
ADOZA
ADXAS
ADZMN
AEFGJ
AEIGN
AEIMD
AENEX
AEUYR
AFBPY
AFFPM
AFGKR
AFWVQ
AFZJQ
AGHNM
AGXDD
AGYGG
AHBTC
AIDQK
AIDYY
AITYG
AIURR
AJXKR
ALMA_UNASSIGNED_HOLDINGS
ALUQN
ALVPJ
AMBMR
AMYDB
ATUGU
AUFTA
AZVAB
BFHJK
BHBCM
BMNLL
BMXJE
BNHUX
BOGZA
BRXPI
CS3
DCZOG
DPXWK
DR2
DRFUL
DRSTM
DU5
EBS
F5P
G-S
GNP
HBH
HGLYW
HHY
HHZ
HZ~
IX1
KQQ
LATKE
LAW
LEEKS
LITHE
LOXES
LUTES
LYRES
MEWTI
MRFUL
MRSTM
MSFUL
MSSTM
MXFUL
MXSTM
MY~
O66
O9-
OIG
P2P
P2W
QRW
R.K
RIWAO
RNS
ROL
RX1
RYL
SUPJJ
V2E
W99
WBKPD
WFSAM
WIH
WIK
WJL
WOHZO
WXSBR
WYISQ
XV2
Y6R
ZZTAW
~S-
31~
53G
AANHP
AASGY
AAYXX
ACBWZ
ACRPL
ACYXJ
ADNMO
AGQPQ
ASPBG
AVWKF
AZFZN
BDRZF
CITATION
EBD
EJD
EMOBN
FEDTE
GODZA
HVGLF
LH4
SV3
NPM
7SR
7U5
8BQ
8FD
JG9
L7M
7X8
5PM
ID FETCH-LOGICAL-c4269-96f432c8b3eae5ad12b8f98548ed3229d73de832dcb53792a69c05d493b1d89c3
IEDL.DBID 24P
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001523633300001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1613-6810
1613-6829
IngestDate Tue Nov 04 02:05:55 EST 2025
Sat Nov 01 14:59:26 EDT 2025
Mon Nov 10 21:41:12 EST 2025
Thu Sep 04 04:52:54 EDT 2025
Thu Nov 27 00:57:50 EST 2025
Fri Aug 22 09:40:11 EDT 2025
IsDoiOpenAccess true
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 33
Keywords convolutional variational autoencoders
spectral imaging
anomaly detection
spectral anomalies
electron energy loss spectroscopy
Language English
License Attribution-NonCommercial-NoDerivs
2025 The Author(s). Small published by Wiley‐VCH GmbH.
This is an open access article under the terms of the http://creativecommons.org/licenses/by-nc-nd/4.0/ License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made.
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c4269-96f432c8b3eae5ad12b8f98548ed3229d73de832dcb53792a69c05d493b1d89c3
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ORCID 0000-0003-4773-6667
0009-0005-4948-7811
OpenAccessLink https://onlinelibrary.wiley.com/doi/abs/10.1002%2Fsmll.202503019
PMID 40619908
PQID 3241874701
PQPubID 1046358
PageCount 8
ParticipantIDs pubmedcentral_primary_oai_pubmedcentral_nih_gov_12372425
proquest_miscellaneous_3227636535
proquest_journals_3241874701
pubmed_primary_40619908
crossref_primary_10_1002_smll_202503019
wiley_primary_10_1002_smll_202503019_SMLL202503019
PublicationCentury 2000
PublicationDate 2025-08-01
PublicationDateYYYYMMDD 2025-08-01
PublicationDate_xml – month: 08
  year: 2025
  text: 2025-08-01
  day: 01
PublicationDecade 2020
PublicationPlace Germany
PublicationPlace_xml – name: Germany
– name: Weinheim
– name: Hoboken
PublicationTitle Small (Weinheim an der Bergstrasse, Germany)
PublicationTitleAlternate Small
PublicationYear 2025
Publisher Wiley Subscription Services, Inc
John Wiley and Sons Inc
Publisher_xml – name: Wiley Subscription Services, Inc
– name: John Wiley and Sons Inc
References 2021; 9
2010; 10
2019; 91
1979; 79
2023; 4
2023; 18
2011
2015; 521
2019; 57
2005; 435
2023; 19
2024; 30
2002; 416
2015; 349
2023; 108
2024; 36
2006; 313
2007; 99
1998; 86
1989; 28
2018; 6
2021; 15
2004; 92
2023; 22
2021; 11
2018; 4
2020; 195
2022
2022; 4
2002; 84
2015; 114
1895; 58
2000; 77
2019
2018
2017
2015
1993; 212
1979; 9
1988; 336
e_1_2_9_30_1
e_1_2_9_31_1
e_1_2_9_34_1
e_1_2_9_10_1
e_1_2_9_35_1
e_1_2_9_13_1
e_1_2_9_32_1
e_1_2_9_12_1
e_1_2_9_33_1
Hui Y. (e_1_2_9_18_1) 2018
e_1_2_9_15_1
e_1_2_9_38_1
e_1_2_9_14_1
e_1_2_9_39_1
Paszke A. (e_1_2_9_40_1) 2019
e_1_2_9_36_1
e_1_2_9_16_1
Lin R. (e_1_2_9_17_1) 2021; 11
e_1_2_9_19_1
Kingma D. P. (e_1_2_9_21_1) 2022
Rombach R. (e_1_2_9_37_1) 2022
e_1_2_9_41_1
e_1_2_9_42_1
e_1_2_9_20_1
e_1_2_9_22_1
e_1_2_9_45_1
e_1_2_9_24_1
e_1_2_9_43_1
e_1_2_9_23_1
e_1_2_9_44_1
e_1_2_9_8_1
e_1_2_9_7_1
e_1_2_9_5_1
e_1_2_9_4_1
e_1_2_9_3_1
e_1_2_9_2_1
e_1_2_9_1_1
e_1_2_9_9_1
e_1_2_9_26_1
Egerton R. (e_1_2_9_6_1)
e_1_2_9_25_1
e_1_2_9_28_1
e_1_2_9_27_1
e_1_2_9_29_1
Li M. D. (e_1_2_9_11_1) 2015; 114
References_xml – year: 2011
– volume: 91
  year: 2019
  publication-title: Rev. Mod. Phys.
– volume: 416
  start-page: 826
  year: 2002
  publication-title: Nature
– volume: 30
  year: 2024
  publication-title: Microsc. Microanal.
– volume: 336
  start-page: 565
  year: 1988
  publication-title: Nature
– volume: 195
  year: 2020
  publication-title: Comput. Vision Image Understanding
– volume: 84
  start-page: 193
  year: 2002
  publication-title: Catal. Lett.
– volume: 11
  year: 2021
  publication-title: Sci. Rep.
– year: 2019
  publication-title: Pytorch: An imperative style, high‐performance deep learning library
– volume: 28
  start-page: 252
  year: 1989
  publication-title: Ultramicroscopy
– volume: 6
  year: 2018
  publication-title: IEEE Access
– volume: 77
  start-page: 3737
  year: 2000
  publication-title: Appl. Phys. Lett.
– year: 2018
  publication-title: arXiv
– volume: 18
  year: 2023
  publication-title: Appl. Surf. Sci. Adv.
– volume: 521
  start-page: 436
  year: 2015
  publication-title: Nature
– volume: 4
  start-page: 1101
  year: 2022
  publication-title: Nat. Mach. Intell.
– volume: 15
  year: 2021
  publication-title: Front. Neurosci.
– volume: 99
  year: 2007
  publication-title: Phys. Rev. Lett.
– volume: 114
  year: 2015
  publication-title: Phys. Rev. Lett.
– volume: 58
  start-page: 240
  year: 1895
  publication-title: Proc. R. Soc. Lond.
– volume: 22
  start-page: 39
  year: 2023
  publication-title: Struct. Health Monit.
– volume: 79
  start-page: 1
  year: 1979
  end-page: 14
– volume: 4
  year: 2023
  publication-title: Mach. Learn.: Sci. Technol.
– volume: 57
  start-page: 6808
  year: 2019
  publication-title: IEEE Trans. Geosci. Remote Sens.
– volume: 4
  year: 2023
  publication-title: Machine Learning: Science and Technology
– volume: 92
  year: 2004
  publication-title: Phys. Rev. Lett.
– volume: 108
  year: 2023
  publication-title: Phys. Rev. B
– volume: 313
  start-page: 504
  year: 2006
  publication-title: Science
– volume: 349
  start-page: 255
  year: 2015
  publication-title: Science
– volume: 19
  year: 2023
  publication-title: Small
– year: 2017
– volume: 435
  start-page: 475
  year: 2005
  publication-title: Nature
– volume: 4
  year: 2018
  publication-title: Adv. Struct. Chem. Imaging
– volume: 9
  start-page: 62
  year: 1979
  publication-title: IEEE Trans. Syst. Man Cybern.
– volume: 36
  start-page: 2743
  year: 2024
  publication-title: Chem. Mater.
– volume: 10
  start-page: 3209
  year: 2010
  publication-title: Nano Lett.
– year: 2022
  publication-title: arXiv
– year: 2015
– volume: 212
  start-page: 185
  year: 1993
  publication-title: Physica C
– volume: 86
  start-page: 2278
  year: 1998
  publication-title: Proc. IEEE
– volume: 9
  year: 2021
  publication-title: IEEE Access
– ident: e_1_2_9_42_1
– ident: e_1_2_9_44_1
  doi: 10.1103/PhysRevB.108.214428
– ident: e_1_2_9_34_1
  doi: 10.1098/rspl.1895.0041
– ident: e_1_2_9_38_1
  doi: 10.1088/2632-2153/acf6a9
– volume-title: Electron Energy Loss Spectroscopy in the Electron Microscope
  ident: e_1_2_9_6_1
– ident: e_1_2_9_8_1
  doi: 10.1016/0304-3991(89)90304-5
– ident: e_1_2_9_10_1
  doi: 10.1103/PhysRevLett.99.047203
– ident: e_1_2_9_9_1
  doi: 10.1016/0921-4534(93)90501-G
– ident: e_1_2_9_14_1
  doi: 10.1016/j.apsadv.2023.100523
– year: 2022
  ident: e_1_2_9_21_1
  publication-title: arXiv
– ident: e_1_2_9_1_1
  doi: 10.1021/nl1022139
– ident: e_1_2_9_3_1
  doi: 10.1038/416826a
– year: 2018
  ident: e_1_2_9_18_1
  publication-title: arXiv
– ident: e_1_2_9_41_1
  doi: 10.1007/978-1-4419-7200-2
– ident: e_1_2_9_27_1
  doi: 10.1177/14759217211073301
– ident: e_1_2_9_4_1
  doi: 10.1023/A:1021428003587
– volume: 114
  year: 2015
  ident: e_1_2_9_11_1
  publication-title: Phys. Rev. Lett.
– ident: e_1_2_9_36_1
– ident: e_1_2_9_43_1
– ident: e_1_2_9_13_1
  doi: 10.1021/acs.chemmater.3c02773
– ident: e_1_2_9_19_1
  doi: 10.1186/s40679-018-0052-y
– ident: e_1_2_9_24_1
  doi: 10.1088/2632-2153/ad073b
– ident: e_1_2_9_26_1
  doi: 10.1038/s41598-021-97668-8
– ident: e_1_2_9_29_1
  doi: 10.3389/fnins.2021.652987
– ident: e_1_2_9_39_1
  doi: 10.1109/ACCESS.2018.2848210
– ident: e_1_2_9_20_1
  doi: 10.1126/science.1127647
– ident: e_1_2_9_28_1
  doi: 10.1016/j.cviu.2020.102920
– ident: e_1_2_9_16_1
  doi: 10.1126/science.aaa8415
– ident: e_1_2_9_35_1
  doi: 10.1109/TSMC.1979.4310076
– year: 2019
  ident: e_1_2_9_40_1
  publication-title: Pytorch: An imperative style, high‐performance deep learning library
– ident: e_1_2_9_15_1
  doi: 10.1038/nature14539
– ident: e_1_2_9_22_1
  doi: 10.1103/RevModPhys.91.045002
– ident: e_1_2_9_31_1
  doi: 10.1093/mam/ozae044.180
– ident: e_1_2_9_32_1
  doi: 10.1109/5.726791
– year: 2022
  ident: e_1_2_9_37_1
  publication-title: arXiv
– ident: e_1_2_9_23_1
  doi: 10.1109/TGRS.2019.2908756
– ident: e_1_2_9_33_1
– ident: e_1_2_9_25_1
  doi: 10.1038/s42256-022-00555-8
– ident: e_1_2_9_7_1
  doi: 10.1103/PhysRevLett.92.095502
– ident: e_1_2_9_30_1
  doi: 10.1002/smll.202205977
– ident: e_1_2_9_2_1
  doi: 10.1038/nature03644
– ident: e_1_2_9_45_1
  doi: 10.1109/ACCESS.2021.3065838
– ident: e_1_2_9_5_1
  doi: 10.1038/336565a0
– volume: 11
  year: 2021
  ident: e_1_2_9_17_1
  publication-title: Sci. Rep.
– ident: e_1_2_9_12_1
  doi: 10.1063/1.1330572
SSID ssj0031247
Score 2.484632
Snippet A 3D Convolutional Variational Autoencoder (3D‐CVAE) is introduced for automated anomaly detection in electron energy‐loss spectroscopy spectrum imaging...
A 3D Convolutional Variational Autoencoder (3D-CVAE) is introduced for automated anomaly detection in electron energy-loss spectroscopy spectrum imaging...
SourceID pubmedcentral
proquest
pubmed
crossref
wiley
SourceType Open Access Repository
Aggregation Database
Index Database
Publisher
StartPage e2503019
SubjectTerms Anomalies
anomaly detection
Automation
convolutional variational autoencoders
Defects
Electron energy
electron energy loss spectroscopy
Image reconstruction
Performance evaluation
Principal components analysis
Robustness
Spectra
spectral anomalies
spectral imaging
Spectrum analysis
Title Robust Spectral Anomaly Detection in EELS Spectral Images via 3D Convolutional Variational Autoencoders
URI https://onlinelibrary.wiley.com/doi/abs/10.1002%2Fsmll.202503019
https://www.ncbi.nlm.nih.gov/pubmed/40619908
https://www.proquest.com/docview/3241874701
https://www.proquest.com/docview/3227636535
https://pubmed.ncbi.nlm.nih.gov/PMC12372425
Volume 21
WOSCitedRecordID wos001523633300001&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: PRVWIB
  databaseName: Wiley Online Library Full Collection 2020
  customDbUrl:
  eissn: 1613-6829
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0031247
  issn: 1613-6810
  databaseCode: DRFUL
  dateStart: 20050101
  isFulltext: true
  titleUrlDefault: https://onlinelibrary.wiley.com
  providerName: Wiley-Blackwell
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1La9wwEB7apIf20PfDbRpUKPRkYkuyLR2X7C4tuCEkTdmbsS25XcjKJfYu9N93xvY6u-RQaC_GRhKWNJqXNPoG4GNeClQk6J1YRVdyilD6ysgQCWK45qVRxnQgrmlydqYWC32-c4u_x4cYN9yIMzp5TQyeF83JLWhos7qmowNU4bhG9X04DEOhKHkDl-dbWSxQe3XpVVBp-YS8tYVtDPjJfvt9tXTH1rwbMrlryna6aP7k_0fxFB4Pdiib9AvnGdyz7jk82kEnfAE_Lupi3bSMUtRTb9jE1av8-jeb2rYL4HJs6dhsll7eVvmyQvnUsM0yZ2LKTmu3GZY2ln1Ht3zYemSTdVsThCaFUb-Eq_ns2-lnf8jL4Jd08dXXcSUFL1UhbG6j3IS8UJVW6PtYg_JBm0QYi5LClEUkEs3zWJdBZKQWRWiULsUrOHC1s2-AWatkWImETAfJg0rFViYmiKJCyMrEgQeftmTJfvXwG1kPtMwzmrpsnDoPjrZUywY2bDK0FinnYBKEHnwYi5GB6FQkd7ZeUx2OMjaOROTB657I46_I2kF1rTxQe-QfKxA4936JW_7sQLrRIkjInfOAd_T_S_ezy69pOn69_ZdG7-AhvffRiUdw0N6s7Xt4UG7aZXNz3PEFPpOFOobD6cX8Kv0DsJwQig
linkProvider Wiley-Blackwell
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV1Lb9QwEB5BQYIeeBYaKGAkJE5RE9tJ7OOq3aoV6aqiBfUWJbEDK3Ud1GRX4t8zk2TTrnpAQhwT24rjeXo8_gbgU14KNCS4O7GKruQUofSVkSESxHDNS6OM6UBc02Q2U5eX-mzIJqS7MD0-xBhwI8no9DUJOAWk929QQ5vFFZ0doA1HJtX34YFEU0OszuXZWhkLNF9dfRW0Wj5Bb61xGwO-vzl-0y7dcTbv5kze9mU7Y3T09D_8xjN4MniibNKzznO4Z90L2L6FT_gSfnyti2XTMipST9NhE1cv8qvf7NC2XQqXY3PHptP0_KbLyQI1VMNW85yJQ3ZQu9XA3Nj2HTfmQ_CRTZZtTSCalEi9A9-OphcHx_5QmcEv6eqrr-NKCl6qQtjcRrkJeaEqrXD3Yw1qCG0SYSzqClMWkUg0z2NdBpGRWhShUboUr2DL1c7uArNWybASCTkPkgeViq1MTBBFhZCViQMPPq_pkv3qATiyHmqZZ7R02bh0HuytyZYNgthk6C9S1cEkCD34ODajCNG5SO5svaQ-HLVsHInIg9c9lcdPkb-DBlt5oDboP3YgeO7NFjf_2cF0o0-Q0IbOA94xwF-mn52fpun49OZfBn2AR8cXp2mWnsy-vIXH9J7C31zvwVZ7vbTv4GG5aufN9ftOSP4A-cER9w
linkToPdf http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV3Nb9MwFH-CDiE4ML4JG2AkJE7REjsf9rFaWzERqmljaLcoiR2otDrTklbiv-e9JM1W7TBp4pjYSRy_b_v59wC-ZIVAQ4LRiZF0JCf3A1fqwEeCaK54oaXWLYhrEs_n8vxcHffZhHQWpsOHGBbcSDJafU0Cbi51eXCNGlovL2jvAG04Mql6CDsBVZIZwc7kZHaWbNSxQAPWVlhBu-US-NYGudHjB9tv2LZMt9zN21mTN73Z1hzNdv_DjzyHZ70vysYd87yAB8a-hKc3EApfwe-TKl_VDaMy9TQcNrbVMrv4yyamaZO4LFtYNp0mp9ddjpaoo2q2XmRMTNhhZdc9e2PbLwzN--VHNl41FcFoUir1azibTX8efnP72gxuQYdfXRWVgeCFzIXJTJhpn-eyVBLjH6NRRygdC21QW-giD0WseBapwgt1oETua6kK8QZGtrLmHTBjZOCXIib3IeBeKSMTxNoLw1wEpY48B75u6JJedhAcaQe2zFOaunSYOgf2N2RLe1GsU_QYqe5g7PkOfB6aUYhoZySzplpRH456NgpF6MDbjsrDp8jjQZMtHZBb9B86EED3dotd_GmButEriCmkc4C3DHDH8NPTH0kyXL2_z0Of4PHxZJYmR_Pve_CEbnfJivswaq5W5gM8KtbNor762EvJP2xOEwM
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=Robust+Spectral+Anomaly+Detection+in+EELS+Spectral+Images+via+3D+Convolutional+Variational+Autoencoders&rft.jtitle=Small+%28Weinheim+an+der+Bergstrasse%2C+Germany%29&rft.au=Sultanov%2C+Seyfal&rft.au=Ayyubi%2C+R+A+W&rft.au=Buban%2C+James+P&rft.au=Klie%2C+Robert+F&rft.date=2025-08-01&rft.eissn=1613-6829&rft.volume=21&rft.issue=33&rft.spage=e2503019&rft_id=info:doi/10.1002%2Fsmll.202503019&rft_id=info%3Apmid%2F40619908&rft.externalDocID=40619908
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1613-6810&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1613-6810&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1613-6810&client=summon