Soft Error Tolerant Convolutional Neural Networks on FPGAs With Ensemble Learning

Convolutional neural networks (CNNs) are widely used in computer vision and natural language processing. Field-programmable gate arrays (FPGAs) are popular accelerators for CNNs. However, if used in critical applications, the reliability of FPGA-based CNNs becomes a priority because FPGAs are prone...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE transactions on very large scale integration (VLSI) systems Vol. 30; no. 3; pp. 291 - 302
Main Authors:	Gao, Zhen, Zhang, Han, Yao, Yi, Xiao, Jiajun, Zeng, Shulin, Ge, Guangjun, Wang, Yu, Ullah, Anees, Reviriego, Pedro
Format:	Journal Article
Language:	English
Published:	New York IEEE 01.03.2022 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:	Artificial neural networks Computer vision Convolution Convolutional neural networks Convolutional neural networks (CNNs) ensemble Ensemble learning fault injection Fault tolerant systems Field programmable gate arrays field-programmable gate array (FPGA) accelerator Natural language processing Neural networks Random access memory Redundancy Reliability Reliability analysis Robustness soft error tolerance Soft errors System reliability Systems analysis
ISSN:	1063-8210, 1557-9999
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	Convolutional neural networks (CNNs) are widely used in computer vision and natural language processing. Field-programmable gate arrays (FPGAs) are popular accelerators for CNNs. However, if used in critical applications, the reliability of FPGA-based CNNs becomes a priority because FPGAs are prone to suffer soft errors. Traditional protection schemes, such as triple modular redundancy (TMR), introduce a large overhead, which is not acceptable in resource-limited platforms. This article proposes to use an ensemble of weak CNNs to build a robust classifier with low cost. To have a group of base CNNs with low complexity and balanced similarity and diversity, residual neural networks (ResNets) with different layers (20/32/44/56) are combined in the ensemble system to replace a single strong ResNet 110. In addition, a robust combiner is designed based on the reliability evaluation of a single ResNet. Single ResNets with different layers and different ensemble schemes are implemented on the FPGA accelerator based on Xilinx Zynq 7000 SoC. The reliability of the ensemble systems is evaluated based on a large-scale fault injection platform and compared with that of the TMR-protected ResNet 110 and ResNet 20. Experiment results show that the proposed ensembles could effectively improve the system reliability when suffering soft errors with an overhead much lower than TMR.
AbstractList	Convolutional neural networks (CNNs) are widely used in computer vision and natural language processing. Field-programmable gate arrays (FPGAs) are popular accelerators for CNNs. However, if used in critical applications, the reliability of FPGA-based CNNs becomes a priority because FPGAs are prone to suffer soft errors. Traditional protection schemes, such as triple modular redundancy (TMR), introduce a large overhead, which is not acceptable in resource-limited platforms. This article proposes to use an ensemble of weak CNNs to build a robust classifier with low cost. To have a group of base CNNs with low complexity and balanced similarity and diversity, residual neural networks (ResNets) with different layers (20/32/44/56) are combined in the ensemble system to replace a single strong ResNet 110. In addition, a robust combiner is designed based on the reliability evaluation of a single ResNet. Single ResNets with different layers and different ensemble schemes are implemented on the FPGA accelerator based on Xilinx Zynq 7000 SoC. The reliability of the ensemble systems is evaluated based on a large-scale fault injection platform and compared with that of the TMR-protected ResNet 110 and ResNet 20. Experiment results show that the proposed ensembles could effectively improve the system reliability when suffering soft errors with an overhead much lower than TMR.
Author	Xiao, Jiajun Gao, Zhen Ge, Guangjun Wang, Yu Ullah, Anees Zhang, Han Yao, Yi Zeng, Shulin Reviriego, Pedro
Author_xml	– sequence: 1 givenname: Zhen orcidid: 0000-0001-9887-1418 surname: Gao fullname: Gao, Zhen email: zgao@tju.edu.cn organization: School of Electrical and Information Engineering, Tianjin University, Tianjin, China – sequence: 2 givenname: Han surname: Zhang fullname: Zhang, Han email: z_han@tju.edu.cn organization: Tianjin International Engineering Institute, Tianjin University, Tianjin, China – sequence: 3 givenname: Yi orcidid: 0000-0003-3582-5726 surname: Yao fullname: Yao, Yi email: yiyao@tju.edu.cn organization: School of Electrical and Information Engineering, Tianjin University, Tianjin, China – sequence: 4 givenname: Jiajun surname: Xiao fullname: Xiao, Jiajun email: destinedone@tju.edu.cn organization: School of Electrical and Information Engineering, Tianjin University, Tianjin, China – sequence: 5 givenname: Shulin surname: Zeng fullname: Zeng, Shulin email: zengsl18@mails.tsinghua.edu.cn organization: School of Electronic Engineering, Tsinghua University, Beijing, China – sequence: 6 givenname: Guangjun surname: Ge fullname: Ge, Guangjun email: geguangjun@126.com organization: School of Electronic Engineering, Tsinghua University, Beijing, China – sequence: 7 givenname: Yu orcidid: 0000-0002-2931-8958 surname: Wang fullname: Wang, Yu email: yu-wang@mail.tsinghua.edu.cn organization: School of Electronic Engineering, Tsinghua University, Beijing, China – sequence: 8 givenname: Anees orcidid: 0000-0002-4770-4967 surname: Ullah fullname: Ullah, Anees email: aneesullah@uetpeshawar.edu.pk organization: Department of Electronics Engineering, University of Engineering and Technology, Peshawar, Abbottabad, Pakistan – sequence: 9 givenname: Pedro orcidid: 0000-0003-2540-5234 surname: Reviriego fullname: Reviriego, Pedro email: revirieg@it.uc3m.es organization: Department of Telematic Engineering, Universidad Carlos III de Madrid, Leganés, Spain
BookMark	eNp9kE1LAzEQhoNUsK3-Ab0EPG_N525yLKWtQvEDqx6XdHeiW9dEk1Tx37u14sGDw8AMzPvODM8A9Zx3gNAxJSNKiT5b3i9uL0aMMDrilCuh6R7qUymLTHfR63qS80wxSg7QIMY1IVQITfro5tbbhKch-ICXvoVgXMIT7959u0mNd6bFl7AJ3yV9-PAcsXd4dj0fR_zQpCc8dRFeVi3gBZjgGvd4iPataSMc_dQhuptNl5PzbHE1v5iMF1nFtEwZLbRVVtS2tlatGAcNCqypLNWyzleS6oIoUTFF67oGIJJLwRU3optVOeF8iE53e1-Df9tATOXab0L3cCxZzmVBiy47FdupquBjDGDL19C8mPBZUlJu0ZXf6MotuvIHXWdSf0xVk8wWRwqmaf-3nuysDQD83tK5EpII_gWCLn51
CODEN	ITCOB4
CitedBy_id	crossref_primary_10_1016_j_cosrev_2024_100682 crossref_primary_10_1109_JSAC_2025_3559150 crossref_primary_10_1109_MDAT_2023_3241116 crossref_primary_10_1145_3716140 crossref_primary_10_3390_electronics13050879 crossref_primary_10_1109_TCASAI_2025_3552735 crossref_primary_10_1109_TNS_2022_3216360 crossref_primary_10_1109_TCAD_2024_3523424 crossref_primary_10_1109_TCAD_2023_3266405 crossref_primary_10_1109_JSAC_2023_3242713 crossref_primary_10_1109_TDMR_2023_3235767 crossref_primary_10_3390_aerospace10010088 crossref_primary_10_1109_TNS_2025_3595388
Cites_doi	10.1109/ACCESS.2017.2742698 10.1109/DFT50435.2020.9250837 10.1109/TPDS.2020.3043449 10.1109/TCAD.2017.2705069 10.1109/TC.2018.2792445 10.1109/DAC.2018.8465834 10.1109/TCAD.2020.2981056 10.1109/TNS.2018.2884460 10.1109/ASAP.2019.00-23 10.1109/TAFFC.2017.2753235 10.1109/ISCA.2016.32 10.1109/TMM.2015.2477042 10.1109/TNS.2020.2983662 10.1109/TMI.2016.2525803 10.1109/CVPR.2016.90 10.1109/AICAS48895.2020.9073925 10.1109/TNS.2009.2033381 10.1109/JBHI.2018.2879834 10.1109/AERO.2008.4526470 10.1109/CVPR.2015.7298594 10.1109/LSP.2017.2758862 10.1201/b12207 10.1109/LATW.2017.7906770 10.1109/ICCE-Asia.2016.7804818 10.1109/MDAT.2019.2915656 10.1145/3460288 10.1007/978-0-387-31069-5 10.1109/AERO.2005.1559551 10.1109/ATS47505.2019.000-8 10.1109/TNS.2021.3050707 10.1109/LATW.2018.8347234 10.1109/TIFS.2017.2766583 10.1109/TCAD.2019.2930577 10.1109/TVLSI.2020.3046075 10.1109/ITC44170.2019.9000149 10.23919/DATE48585.2020.9116571 10.1109/CCWC.2018.8301749 10.1109/12.59860 10.1109/TVLSI.2019.2939726 10.1007/978-1-4419-6715-2 10.1109/DSN-W.2017.47 10.1109/TNS.2018.2812719 10.1109/SiPS.2017.8110024
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022
DBID	97E RIA RIE AAYXX CITATION 7SP 8FD L7M
DOI	10.1109/TVLSI.2021.3138491
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef Electronics & Communications Abstracts Technology Research Database Advanced Technologies Database with Aerospace
DatabaseTitle	CrossRef Technology Research Database Advanced Technologies Database with Aerospace Electronics & Communications Abstracts
DatabaseTitleList	Technology Research Database
Database_xml	– sequence: 1 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering
EISSN	1557-9999
EndPage	302
ExternalDocumentID	10_1109_TVLSI_2021_3138491 9684504
Genre	orig-research
GrantInformation_xml	– fundername: Go2Edge Network grantid: RED2018-102585-T – fundername: ACHILLES Project grantid: PID2019-104207RB-I00 – fundername: National Natural Science Foundation of China (NSFC) grantid: 62171313 funderid: 10.13039/501100001809 – fundername: Department of Research and Innovation of Madrid Regional Authority with the EMPATIA-CM Research Project grantid: Y2018/TCS-5046 – fundername: NSFC Joint Foundation grantid: 20200509 funderid: 10.13039/501100001809
GroupedDBID	-~X .DC 0R~ 29I 3EH 4.4 5GY 5VS 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABFSI ABQJQ ABVLG ACGFS ACIWK AENEX AETIX AGQYO AGSQL AHBIQ AI. AIBXA AKJIK AKQYR ALLEH ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 E.L EBS EJD HZ~ H~9 ICLAB IEDLZ IFIPE IFJZH IPLJI JAVBF LAI M43 O9- OCL P2P RIA RIE RNS TN5 VH1 AAYXX CITATION 7SP 8FD L7M
ID	FETCH-LOGICAL-c295t-179f8f4dfdff8b23e9e8efacf195d6b5197084c281dddee05354383a46b5c6033
IEDL.DBID	RIE
ISICitedReferencesCount	21
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000745446700001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	1063-8210
IngestDate	Mon Jun 30 10:26:27 EDT 2025 Tue Nov 18 22:18:54 EST 2025 Sat Nov 29 03:36:18 EST 2025 Wed Aug 27 02:49:31 EDT 2025
IsPeerReviewed	true
IsScholarly	true
Issue	3
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c295t-179f8f4dfdff8b23e9e8efacf195d6b5197084c281dddee05354383a46b5c6033
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
ORCID	0000-0002-2931-8958 0000-0003-3582-5726 0000-0001-9887-1418 0000-0002-4770-4967 0000-0003-2540-5234
PQID	2635717717
PQPubID	85424
PageCount	12
ParticipantIDs	proquest_journals_2635717717 crossref_citationtrail_10_1109_TVLSI_2021_3138491 crossref_primary_10_1109_TVLSI_2021_3138491 ieee_primary_9684504
PublicationCentury	2000
PublicationDate	2022-03-01
PublicationDateYYYYMMDD	2022-03-01
PublicationDate_xml	– month: 03 year: 2022 text: 2022-03-01 day: 01
PublicationDecade	2020
PublicationPlace	New York
PublicationPlace_xml	– name: New York
PublicationTitle	IEEE transactions on very large scale integration (VLSI) systems
PublicationTitleAbbrev	TVLSI
PublicationYear	2022
Publisher	IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml	– name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References	ref13 ref15 ref14 ref52 Simonyan (ref1) ref10 ref17 ref16 ref19 ref51 ref50 Caffrey (ref18) ref46 ref45 ref47 ref42 ref41 ref44 ref43 ref7 ref9 ref4 Lee (ref49) 2014 ref3 ref5 ref35 ref34 ref37 ref36 ref31 ref30 ref33 ref32 ref2 ref39 ref38 Li (ref8) ref24 ref23 ref26 ref25 ref20 Carmichael (ref11) 2006 ref22 ref21 (ref6) 2011 (ref40) 2020 ref28 ref27 ref29 Ioffe (ref48) Jonathan (ref12)
References_xml	– ident: ref15 doi: 10.1109/ACCESS.2017.2742698 – ident: ref16 doi: 10.1109/DFT50435.2020.9250837 – ident: ref32 doi: 10.1109/TPDS.2020.3043449 – ident: ref4 doi: 10.1109/TCAD.2017.2705069 – ident: ref51 doi: 10.1109/TC.2018.2792445 – ident: ref23 doi: 10.1109/DAC.2018.8465834 – volume-title: Proc. ICML ident: ref48 article-title: Batch normalization: Accelerating deep network training by reducing internal covariate shift – volume-title: Road Vehicles—Functional Safety year: 2011 ident: ref6 – ident: ref33 doi: 10.1109/TCAD.2020.2981056 – ident: ref26 doi: 10.1109/TNS.2018.2884460 – ident: ref34 doi: 10.1109/ASAP.2019.00-23 – ident: ref47 doi: 10.1109/TAFFC.2017.2753235 – ident: ref21 doi: 10.1109/ISCA.2016.32 – ident: ref46 doi: 10.1109/TMM.2015.2477042 – ident: ref28 doi: 10.1109/TNS.2020.2983662 – ident: ref42 doi: 10.1109/TMI.2016.2525803 – ident: ref3 doi: 10.1109/CVPR.2016.90 – ident: ref38 doi: 10.1109/AICAS48895.2020.9073925 – ident: ref17 doi: 10.1109/TNS.2009.2033381 – ident: ref43 doi: 10.1109/JBHI.2018.2879834 – ident: ref50 doi: 10.1109/AERO.2008.4526470 – ident: ref2 doi: 10.1109/CVPR.2015.7298594 – ident: ref44 doi: 10.1109/LSP.2017.2758862 – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref1 article-title: Very deep convolutional networks for large-scale image recognition – ident: ref41 doi: 10.1201/b12207 – start-page: 1 volume-title: Proc. Mil. Aerosp. Appl. Program. Devices Technol. Conf. (MAPLD) ident: ref18 article-title: Single-event upsets in SRAM FPGAs – ident: ref25 doi: 10.1109/LATW.2017.7906770 – ident: ref20 doi: 10.1109/ICCE-Asia.2016.7804818 – ident: ref36 doi: 10.1109/MDAT.2019.2915656 – ident: ref35 doi: 10.1145/3460288 – ident: ref9 doi: 10.1007/978-0-387-31069-5 – ident: ref13 doi: 10.1109/AERO.2005.1559551 – ident: ref22 doi: 10.1109/ATS47505.2019.000-8 – ident: ref29 doi: 10.1109/TNS.2021.3050707 – volume-title: Triple module redundancy design techniques for virtex FPGAs year: 2006 ident: ref11 – ident: ref27 doi: 10.1109/LATW.2018.8347234 – ident: ref45 doi: 10.1109/TIFS.2017.2766583 – ident: ref39 doi: 10.1109/TCAD.2019.2930577 – ident: ref10 doi: 10.1109/TVLSI.2020.3046075 – ident: ref31 doi: 10.1109/ITC44170.2019.9000149 – ident: ref24 doi: 10.23919/DATE48585.2020.9116571 – ident: ref19 doi: 10.1109/CCWC.2018.8301749 – start-page: 1 volume-title: Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal. (SC) ident: ref8 article-title: Understanding error propagation in deep learning neural network (DNN) accelerators and applications – start-page: 249 volume-title: Proc. FPGA ident: ref12 article-title: Voter insertion algorithms for FPGA designs using triple modular redundancy – ident: ref14 doi: 10.1109/12.59860 – ident: ref5 doi: 10.1109/TVLSI.2019.2939726 – ident: ref7 doi: 10.1007/978-1-4419-6715-2 – ident: ref30 doi: 10.1109/DSN-W.2017.47 – volume-title: Zynq DPU V3.2-Product Guide, PG338 (V3.2) year: 2020 ident: ref40 – ident: ref52 doi: 10.1109/TNS.2018.2812719 – year: 2014 ident: ref49 article-title: Deeply-supervised nets publication-title: arXiv:1409.5185 – ident: ref37 doi: 10.1109/SiPS.2017.8110024
SSID	ssj0014490
Score	2.4504898
Snippet	Convolutional neural networks (CNNs) are widely used in computer vision and natural language processing. Field-programmable gate arrays (FPGAs) are popular...
SourceID	proquest crossref ieee
SourceType	Aggregation Database Enrichment Source Index Database Publisher
StartPage	291
SubjectTerms	Artificial neural networks Computer vision Convolution Convolutional neural networks Convolutional neural networks (CNNs) ensemble Ensemble learning fault injection Fault tolerant systems Field programmable gate arrays field-programmable gate array (FPGA) accelerator Natural language processing Neural networks Random access memory Redundancy Reliability Reliability analysis Robustness soft error tolerance Soft errors System reliability Systems analysis
Title	Soft Error Tolerant Convolutional Neural Networks on FPGAs With Ensemble Learning
URI	https://ieeexplore.ieee.org/document/9684504 https://www.proquest.com/docview/2635717717
Volume	30
WOSCitedRecordID	wos000745446700001&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: PRVIEE databaseName: IEEE Electronic Library (IEL) customDbUrl: eissn: 1557-9999 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0014490 issn: 1063-8210 databaseCode: RIE dateStart: 19930101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3fS8MwEA5z-KAP_pridEoefNO6pk3b5HGMTYUxJpu6t5KmiQ5mK223v98k7YqiCEKhpUmg3Jfe5ZK77wC4Qoz7kU-lFceBsLD0Y4tFLLKYTQlzJEIRMzyzo2A8JvM5nTTATZ0LI4QwwWfiVj-as_w45Su9VdalPsGeJv_cCoKgzNWqTwwwpiXzgO9aRPkxmwQZm3Znz6Ppg3IFHaQ8VJdgir4ZIVNV5YcqNvZluP-_LzsAe9U6EvZK4A9BQyRHYPcLu2ALPE6VioWDLEszOEuXQlmlAvbTZF3NNjVcU3OYm4kFz2GawOHkrpfDl0XxBgdJLt6jpYAVCevrMXgaDmb9e6uqoGBxh3qF5h6VROJYxlKSyHEFFURIxiWiXuxHOmnVJpg7atGq1JyuEuFp6lKGVRv3bdc9Ac0kTcQpgA5SoEqfYWyrZo8ym4koRly9E4R7dhugjUhDXtGL6yoXy9C4GTYNDQyhhiGsYGiD63rMR0mu8WfvlhZ83bOSeRt0NsiF1f-Xh5piRzmq6jr7fdQ52HF0IoOJJuuAZpGtxAXY5utikWeXZmp9AlU1zM8
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3dS8MwED9kCuqD3-L8zINvWte0aZc8jrGpOIfi_HgraZroYLbSdvv7TbJuKIogFFqaBMr90rtccvc7gFPMRRiHTDlJ0pQOUWHi8JjHDncZ5Z7COOaWZ7bX7Pfpywu7W4DzeS6MlNIGn8kL82jP8pNMjM1WWYOFlASG_HMxIMTD02yt-ZkBIWzKPRD6DtWezCxFxmWNwVPv4Vo7gx7WPqpPCcPfzJCtq_JDGVsL013_37dtwFq1kkStKfSbsCDTLVj9wi-4DfcPWsmiTp5nORpkI6ntUonaWTqp5psebsg57M1GgxcoS1H37rJVoOdh-YY6aSHf45FEFQ3r6w48djuD9pVT1VBwhMeC0rCPKqpIohKlaOz5kkkqFRcKsyAJY5O26lIiPL1s1YrO1IkIDHkpJ7pNhK7v70ItzVK5B8jDGlYVckJc3Rww7nIZJ1jod5KKwK0Dnok0EhXBuKlzMYqso-GyyMIQGRiiCoY6nM3HfEzpNf7svW0EP-9ZybwOhzPkouoPLCJDsqNdVX3t_z7qBJavBre9qHfdvzmAFc-kNdjYskOolflYHsGSmJTDIj-20-wTJzXQFg
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=Soft+Error+Tolerant+Convolutional+Neural+Networks+on+FPGAs+With+Ensemble+Learning&rft.jtitle=IEEE+transactions+on+very+large+scale+integration+%28VLSI%29+systems&rft.au=Gao%2C+Zhen&rft.au=Zhang%2C+Han&rft.au=Yao%2C+Yi&rft.au=Xiao%2C+Jiajun&rft.date=2022-03-01&rft.pub=IEEE&rft.issn=1063-8210&rft.volume=30&rft.issue=3&rft.spage=291&rft.epage=302&rft_id=info:doi/10.1109%2FTVLSI.2021.3138491&rft.externalDocID=9684504
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1063-8210&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1063-8210&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1063-8210&client=summon