MobileCount: An efficient encoder-decoder framework for real-time crowd counting

In this work, we propose a computation-efficient encoder-decoder architecture, named MobileCount, which is specifically designed for high-accuracy real-time crowd counting on mobile or embedded devices with limited computation resources. For the encoder part, MobileNetV2 is tailored in order to sign...

Full description

Saved in:
Bibliographic Details
Published in:Neurocomputing (Amsterdam) Vol. 407; pp. 292 - 299
Main Authors: Wang, Peng, Gao, Chenyu, Wang, Yang, Li, Hui, Gao, Ye
Format: Journal Article
Language:English
Published: Elsevier B.V 24.09.2020
Subjects:
ISSN:0925-2312, 1872-8286
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract In this work, we propose a computation-efficient encoder-decoder architecture, named MobileCount, which is specifically designed for high-accuracy real-time crowd counting on mobile or embedded devices with limited computation resources. For the encoder part, MobileNetV2 is tailored in order to significantly reduce FLOPs at a little cost of performance drop, which has 4 bottleneck blocks preceded by a max pooling layer of stride 2. The design of decoder is motivated by Light-weight RefineNet, which further boosts counting performance with only a 10% increase of FLOPs. In comparison with state-of-the-arts, our proposed network is able to achieve comparable counting performance with 1/10 FLOPs on a number of benchmarks. At last, we propose a multi-layer knowledge distillation method to further boost the performance of MobileCount without increasing its FLOPs.
AbstractList In this work, we propose a computation-efficient encoder-decoder architecture, named MobileCount, which is specifically designed for high-accuracy real-time crowd counting on mobile or embedded devices with limited computation resources. For the encoder part, MobileNetV2 is tailored in order to significantly reduce FLOPs at a little cost of performance drop, which has 4 bottleneck blocks preceded by a max pooling layer of stride 2. The design of decoder is motivated by Light-weight RefineNet, which further boosts counting performance with only a 10% increase of FLOPs. In comparison with state-of-the-arts, our proposed network is able to achieve comparable counting performance with 1/10 FLOPs on a number of benchmarks. At last, we propose a multi-layer knowledge distillation method to further boost the performance of MobileCount without increasing its FLOPs.
Author Wang, Yang
Wang, Peng
Li, Hui
Gao, Ye
Gao, Chenyu
Author_xml – sequence: 1
  givenname: Peng
  surname: Wang
  fullname: Wang, Peng
  organization: School of Computer Science, Northwestern Polytechnical University, Xi’an, China
– sequence: 2
  givenname: Chenyu
  surname: Gao
  fullname: Gao, Chenyu
  organization: School of Software, Northwestern Polytechnical University, Xi’an, China
– sequence: 3
  givenname: Yang
  surname: Wang
  fullname: Wang, Yang
  organization: School of Computer Science, Northwestern Polytechnical University, Xi’an, China
– sequence: 4
  givenname: Hui
  surname: Li
  fullname: Li, Hui
  email: hui.li02@adelaide.edu.au
  organization: University of Adelaide, Australia
– sequence: 5
  givenname: Ye
  surname: Gao
  fullname: Gao, Ye
  organization: School of Computer Science, Northwestern Polytechnical University, Xi’an, China
BookMark eNqFkM1KxDAUhYOM4Dj6Bi7yAq1J2qTpLIRh8A9GdDH7kKY3krFNJM04-Pa2jisXCgfO6juX-52jmQ8eELqiJKeEiutd7mFvQp8zwkhO-BhxguZUViyTTIoZmpOa8YwVlJ2h82HYEUIryuo5enkKjetgHfY-LfHKY7DWGQc-YfAmtBCzFr4b26h7OIT4hm2IOILusuR6wCaGQ4vNtOD86wU6tbob4PKnF2h7d7tdP2Sb5_vH9WqTmYKIlMmayoa03ILVDdekEURzQ1lhiSiZtmVFoeRac06ZlI0B0RTE1rwpRcUlLRZoeZwdrw9DBKuMSzq54FPUrlOUqEmN2qmjGjWpUYSPESNc_oLfo-t1_PwPuzliMP714SCqYTJloHURTFJtcH8PfAFA2oMU
CitedBy_id crossref_primary_10_1007_s11554_024_01531_8
crossref_primary_10_1007_s00530_024_01318_8
crossref_primary_10_3390_electronics13040723
crossref_primary_10_1016_j_neucom_2023_01_059
crossref_primary_10_1111_exsy_70098
crossref_primary_10_1016_j_future_2023_05_013
crossref_primary_10_1007_s10845_022_02024_w
crossref_primary_10_1016_j_cosrev_2024_100645
crossref_primary_10_1016_j_neucom_2023_126914
crossref_primary_10_1016_j_asoc_2024_111366
crossref_primary_10_1016_j_knosys_2023_110541
crossref_primary_10_1002_aisy_202300185
crossref_primary_10_1061_JTEPBS_TEENG_8944
crossref_primary_10_1109_ACCESS_2021_3078481
crossref_primary_10_1109_TIP_2023_3343609
crossref_primary_10_1109_TAES_2021_3087821
crossref_primary_10_1016_j_neucom_2022_08_037
crossref_primary_10_1109_JIOT_2025_3542534
crossref_primary_10_4218_etrij_2023_0426
crossref_primary_10_3390_rs14246363
crossref_primary_10_1109_TCSVT_2022_3171235
crossref_primary_10_1016_j_displa_2024_102686
crossref_primary_10_1109_ACCESS_2022_3171328
crossref_primary_10_1016_j_neucom_2021_11_099
crossref_primary_10_1007_s00371_024_03572_3
crossref_primary_10_1016_j_engappai_2022_105698
crossref_primary_10_1109_TGRS_2023_3238185
crossref_primary_10_32604_cmc_2023_035974
crossref_primary_10_3390_s21103464
crossref_primary_10_1088_1361_6501_adea12
crossref_primary_10_1080_21681163_2023_2188971
crossref_primary_10_1088_1361_6501_ada6f1
crossref_primary_10_1109_ACCESS_2022_3144607
crossref_primary_10_1016_j_eswa_2022_116662
crossref_primary_10_1007_s10489_023_04499_3
crossref_primary_10_1016_j_neucom_2023_126662
crossref_primary_10_3233_JIFS_224081
crossref_primary_10_32604_cmc_2024_048928
crossref_primary_10_1016_j_neucom_2021_02_047
crossref_primary_10_1007_s00500_021_05993_x
crossref_primary_10_1016_j_eswa_2023_122069
crossref_primary_10_1007_s10586_025_05226_y
crossref_primary_10_1117_1_JEI_33_4_043002
crossref_primary_10_1109_TGRS_2021_3125249
crossref_primary_10_1016_j_neucom_2021_06_086
crossref_primary_10_1109_TCSVT_2024_3469933
crossref_primary_10_1007_s00371_023_03099_z
crossref_primary_10_1109_TIM_2025_3588967
crossref_primary_10_1007_s11831_022_09772_1
Cites_doi 10.1145/1150402.1150464
10.1109/TITS.2017.2750080
10.1007/s11263-005-6644-8
ContentType Journal Article
Copyright 2020 Elsevier B.V.
Copyright_xml – notice: 2020 Elsevier B.V.
DBID AAYXX
CITATION
DOI 10.1016/j.neucom.2020.05.056
DatabaseName CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Computer Science
EISSN 1872-8286
EndPage 299
ExternalDocumentID 10_1016_j_neucom_2020_05_056
S0925231220308912
GroupedDBID ---
--K
--M
.DC
.~1
0R~
123
1B1
1~.
1~5
4.4
457
4G.
53G
5VS
7-5
71M
8P~
9JM
9JN
AABNK
AACTN
AADPK
AAEDT
AAEDW
AAIAV
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAXLA
AAXUO
AAYFN
ABBOA
ABCQJ
ABFNM
ABJNI
ABMAC
ABYKQ
ACDAQ
ACGFS
ACRLP
ACZNC
ADBBV
ADEZE
AEBSH
AEKER
AENEX
AFKWA
AFTJW
AFXIZ
AGHFR
AGUBO
AGWIK
AGYEJ
AHHHB
AHZHX
AIALX
AIEXJ
AIKHN
AITUG
AJOXV
ALMA_UNASSIGNED_HOLDINGS
AMFUW
AMRAJ
AOUOD
AXJTR
BKOJK
BLXMC
CS3
DU5
EBS
EFJIC
EFLBG
EO8
EO9
EP2
EP3
F5P
FDB
FIRID
FNPLU
FYGXN
G-Q
GBLVA
GBOLZ
IHE
J1W
KOM
LG9
M41
MO0
MOBAO
N9A
O-L
O9-
OAUVE
OZT
P-8
P-9
P2P
PC.
Q38
ROL
RPZ
SDF
SDG
SDP
SES
SPC
SPCBC
SSN
SSV
SSZ
T5K
ZMT
~G-
29N
9DU
AAQXK
AATTM
AAXKI
AAYWO
AAYXX
ABWVN
ABXDB
ACLOT
ACNNM
ACRPL
ACVFH
ADCNI
ADJOM
ADMUD
ADNMO
AEIPS
AEUPX
AFJKZ
AFPUW
AGQPQ
AIGII
AIIUN
AKBMS
AKRWK
AKYEP
ANKPU
APXCP
ASPBG
AVWKF
AZFZN
CITATION
EFKBS
EJD
FEDTE
FGOYB
HLZ
HVGLF
HZ~
R2-
SBC
SEW
WUQ
XPP
~HD
ID FETCH-LOGICAL-c306t-8918b0d5fefab5a0b60a5c123f0642af471e45aa551288bce6b30f95b4675813
ISICitedReferencesCount 61
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000555461000011&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0925-2312
IngestDate Tue Nov 18 22:35:27 EST 2025
Sat Nov 29 07:10:58 EST 2025
Fri Feb 23 02:46:21 EST 2024
IsPeerReviewed true
IsScholarly true
Keywords Fully convolutional networks
Light-weight Neural Networks
Knowledge distillation
Crowd counting
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c306t-8918b0d5fefab5a0b60a5c123f0642af471e45aa551288bce6b30f95b4675813
PageCount 8
ParticipantIDs crossref_citationtrail_10_1016_j_neucom_2020_05_056
crossref_primary_10_1016_j_neucom_2020_05_056
elsevier_sciencedirect_doi_10_1016_j_neucom_2020_05_056
PublicationCentury 2000
PublicationDate 2020-09-24
PublicationDateYYYYMMDD 2020-09-24
PublicationDate_xml – month: 09
  year: 2020
  text: 2020-09-24
  day: 24
PublicationDecade 2020
PublicationTitle Neurocomputing (Amsterdam)
PublicationYear 2020
Publisher Elsevier B.V
Publisher_xml – name: Elsevier B.V
References A.G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, H. Adam, Mobilenets: Efficient convolutional neural networks for mobile vision applications, arXiv preprint arXiv:1704.04861.
A. Krizhevsky, I. Sutskever, G.E. Hinton, Imagenet classification with deep convolutional neural networks, in: Proc. Advances in Neural Inf. Process. Syst., 2012.
V. Nekrasov, C. Shen, I. Reid, Light-weight refinenet for real-time semantic segmentations, arXiv preprint arXiv:1810.03272.
Hubara, Courbariaux, Soudry, El-Yaniv, Bengio (b0175) 2017; 18
Lan, Zhu, Gong (b0210) 2018
A. Paszke, A. Chaurasia, S. Kim, E. Culurciello, Enet: A deep neural network architecture for real-time semantic segmentation, arXiv preprint arXiv:1606.02147.
Redmon, Divvala, Girshick, Farhadi (b0125) 2016
S. Han, J. Pool, J. Tran, W. Dally, Learning both weights and connections for efficient neural network, in: Proc. Advances in Neural Inf. Process. Syst., 2015, pp. 1135–1143.
Zhang, Li, Wang, Yang (b0055) 2015
Ren, He, Girshick, Sun (b0120) 2015
Cao, Wang, Zhao, Su (b0235) 2018
V. Nekrasov, H. Chen, C. Shen, I. Reid, Fast neural architecture search of compact semantic segmentation models via auxiliary cells, arXiv preprint arXiv:1810.10804.
Lin, Dollár, Girshick, He, Hariharan, Belongie (b0230) 2017
Sandler, Howard, Zhu, Zhmoginov, Chen (b0100) 2018
Romera, Alvarez, Bergasa, Arroyo (b0110) 2018; 19
Shi, Zhang, Liu, Cao, Ye, Cheng, Zheng (b0140) 2018
Sindagi, Patel (b0135) 2017
Zhang, Zhou, Chen, Gao, Ma (b0065) 2016
J. Gao, W. Lin, B. Zhao, D. Wang, C. Gao, J. Wen, C-3-Framework: An open-source pytorch code for crowd counting, arXiv preprint arXiv:1907.02724.
He, Zhang, Ren, Sun (b0025) 2016
Veit, Belongie (b0170) 2018
M. Tan, B. Chen, R. Pang, V. Vasudevan, Q.V. Le, Mnasnet: Platform-aware neural architecture search for mobile, arXiv preprint arXiv:1807.11626.
L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, A.L. Yuille, Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs., IEEE Trans. Pattern Anal. Mach. Intell.
C. Bucilu, R. Caruana, A. Niculescu-Mizil, Model compression, in: Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining, ACM, 2006, pp. 535–541.
Girshick, Donahue, Darrell, Malik (b0005) 2014
Simonyan, Zisserman (b0020) 2015
Ba, Caruana (b0185) 2014
Sam, Sajjan, Babu, Srinivasan (b0075) 2018
Long, Shelhamer, Darrell (b0030) 2015
Zhang, Xiang, Hospedales, Lu (b0205) 2018
P. Viola, M.J. Jones, D. Snow, Detecting pedestrians using patterns of motion and appearance, in: Int. J. Comput. Vision, 2005.
Chan, Vasconcelos (b0050) 2009
Cai, Vasconcelos (b0010) 2018
Zhang, Zhu, Ye (b0200) 2019
Lin, Shen, Van Den Hengel, Reid (b0035) 2016
Lin, Milan, Shen, Reid (b0145) 2017
Ma, Zhang, Zheng, Sun (b0160) 2018
Boominathan, Kruthiventi, Babu (b0060) 2016
A. Romero, N. Ballas, S.E. Kahou, A. Chassang, C. Gatta, Y. Bengio, Fitnets: Hints for thin deep nets, arXiv preprint arXiv:1412.6550.
Sam, Surya, Babu (b0070) 2017
G. Hinton, O. Vinyals, J. Dean, Distilling the knowledge in a neural network, arXiv preprint arXiv:1503.02531.
Zhao, Li, Abu Alsheikh, Tian, Zhao, Torralba, Katabi (b0215) 2018
Zhang, Zhou, Lin, Sun (b0095) 2018
Li, Zhang, Chen (b0080) 2018
F.N. Iandola, S. Han, M.W. Moskewicz, K. Ashraf, W.J. Dally, K. Keutzer, Squeezenet: Alexnet-level accuracy with 50x fewer parameters and< 0.5 mb model size, arXiv preprint arXiv:1602.07360.
Idrees, Tayyab, Athrey, Zhang, Al-Maadeed, Rajpoot, Shah (b0085) 2018
Idrees, Saleemi, Seibert, Shah (b0225) 2013
Sam (10.1016/j.neucom.2020.05.056_b0075) 2018
Zhang (10.1016/j.neucom.2020.05.056_b0095) 2018
Cao (10.1016/j.neucom.2020.05.056_b0235) 2018
Zhang (10.1016/j.neucom.2020.05.056_b0065) 2016
Zhang (10.1016/j.neucom.2020.05.056_b0055) 2015
10.1016/j.neucom.2020.05.056_b0130
Lin (10.1016/j.neucom.2020.05.056_b0035) 2016
10.1016/j.neucom.2020.05.056_b0155
10.1016/j.neucom.2020.05.056_b0115
10.1016/j.neucom.2020.05.056_b0015
10.1016/j.neucom.2020.05.056_b0090
Lin (10.1016/j.neucom.2020.05.056_b0230) 2017
10.1016/j.neucom.2020.05.056_b0190
Romera (10.1016/j.neucom.2020.05.056_b0110) 2018; 19
10.1016/j.neucom.2020.05.056_b0195
10.1016/j.neucom.2020.05.056_b0150
Zhao (10.1016/j.neucom.2020.05.056_b0215) 2018
Girshick (10.1016/j.neucom.2020.05.056_b0005) 2014
Simonyan (10.1016/j.neucom.2020.05.056_b0020) 2015
Lin (10.1016/j.neucom.2020.05.056_b0145) 2017
Sandler (10.1016/j.neucom.2020.05.056_b0100) 2018
Sam (10.1016/j.neucom.2020.05.056_b0070) 2017
Li (10.1016/j.neucom.2020.05.056_b0080) 2018
10.1016/j.neucom.2020.05.056_b0105
Shi (10.1016/j.neucom.2020.05.056_b0140) 2018
Lan (10.1016/j.neucom.2020.05.056_b0210) 2018
Sindagi (10.1016/j.neucom.2020.05.056_b0135) 2017
Chan (10.1016/j.neucom.2020.05.056_b0050) 2009
Ba (10.1016/j.neucom.2020.05.056_b0185) 2014
Idrees (10.1016/j.neucom.2020.05.056_b0225) 2013
Ren (10.1016/j.neucom.2020.05.056_b0120) 2015
10.1016/j.neucom.2020.05.056_b0045
Ma (10.1016/j.neucom.2020.05.056_b0160) 2018
10.1016/j.neucom.2020.05.056_b0165
10.1016/j.neucom.2020.05.056_b0220
10.1016/j.neucom.2020.05.056_b0180
Redmon (10.1016/j.neucom.2020.05.056_b0125) 2016
Zhang (10.1016/j.neucom.2020.05.056_b0205) 2018
Boominathan (10.1016/j.neucom.2020.05.056_b0060) 2016
Veit (10.1016/j.neucom.2020.05.056_b0170) 2018
Long (10.1016/j.neucom.2020.05.056_b0030) 2015
10.1016/j.neucom.2020.05.056_b0040
Idrees (10.1016/j.neucom.2020.05.056_b0085) 2018
He (10.1016/j.neucom.2020.05.056_b0025) 2016
Hubara (10.1016/j.neucom.2020.05.056_b0175) 2017; 18
Cai (10.1016/j.neucom.2020.05.056_b0010) 2018
Zhang (10.1016/j.neucom.2020.05.056_b0200) 2019
References_xml – start-page: 6848
  year: 2018
  end-page: 6856
  ident: b0095
  article-title: Shufflenet: an extremely efficient convolutional neural network for mobile devices
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2018
  ident: b0140
  article-title: Crowd counting with deep negative correlation learning
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn
– start-page: 1925
  year: 2017
  end-page: 1934
  ident: b0145
  article-title: Refinenet: multi-path refinement networks for high-resolution semantic segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 580
  year: 2014
  end-page: 587
  ident: b0005
  article-title: Rich feature hierarchies for accurate object detection and semantic segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 91
  year: 2015
  end-page: 99
  ident: b0120
  article-title: Faster r-cnn: Towards real-time object detection with region proposal networks
  publication-title: Adv. Neural Inform. Process. Syst.
– reference: M. Tan, B. Chen, R. Pang, V. Vasudevan, Q.V. Le, Mnasnet: Platform-aware neural architecture search for mobile, arXiv preprint arXiv:1807.11626.
– start-page: 6154
  year: 2018
  end-page: 6162
  ident: b0010
  article-title: Cascade r-cnn: delving into high quality object detection
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 116
  year: 2018
  end-page: 131
  ident: b0160
  article-title: Shufflenet v2: practical guidelines for efficient cnn architecture design
  publication-title: Proc. Eur. Conf. Comp. Vis.
– start-page: 7528
  year: 2018
  end-page: 7538
  ident: b0210
  article-title: Knowledge distillation by on-the-fly native ensemble
  publication-title: Proceedings of the 32nd International Conference on Neural Information Processing Systems, Curran Associates Inc
– reference: A. Krizhevsky, I. Sutskever, G.E. Hinton, Imagenet classification with deep convolutional neural networks, in: Proc. Advances in Neural Inf. Process. Syst., 2012.
– start-page: 2654
  year: 2014
  end-page: 2662
  ident: b0185
  article-title: Do deep nets really need to be deep?
  publication-title: Adv. Neural Inform. Process. Syst.
– start-page: 770
  year: 2016
  end-page: 778
  ident: b0025
  article-title: Deep residual learning for image recognition
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– reference: A. Paszke, A. Chaurasia, S. Kim, E. Culurciello, Enet: A deep neural network architecture for real-time semantic segmentation, arXiv preprint arXiv:1606.02147.
– reference: A. Romero, N. Ballas, S.E. Kahou, A. Chassang, C. Gatta, Y. Bengio, Fitnets: Hints for thin deep nets, arXiv preprint arXiv:1412.6550.
– reference: V. Nekrasov, C. Shen, I. Reid, Light-weight refinenet for real-time semantic segmentations, arXiv preprint arXiv:1810.03272.
– year: 2013
  ident: b0225
  article-title: Multi-source multi-scale counting in extremely dense crowd images
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– volume: 18
  start-page: 6869
  year: 2017
  end-page: 6898
  ident: b0175
  article-title: Quantized neural networks: training neural networks with low precision weights and activations
  publication-title: J. Mach. Learn. Res.
– start-page: 7356
  year: 2018
  end-page: 7365
  ident: b0215
  article-title: Through-wall human pose estimation using radio signals
  publication-title: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition
– start-page: 3
  year: 2018
  end-page: 18
  ident: b0170
  article-title: Convolutional networks with adaptive inference graphs
  publication-title: Proc. Eur. Conf. Comp. Vis.
– start-page: 2117
  year: 2017
  end-page: 2125
  ident: b0230
  article-title: Feature pyramid networks for object detection
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2009
  ident: b0050
  article-title: Bayesian poisson regression for crowd counting
  publication-title: Proc. IEEE Int. Conf. Comp. Vis.
– start-page: 4320
  year: 2018
  end-page: 4328
  ident: b0205
  article-title: Deep mutual learning
  publication-title: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition
– year: 2017
  ident: b0070
  article-title: Switching convolutional neural network for crowd counting
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– reference: G. Hinton, O. Vinyals, J. Dean, Distilling the knowledge in a neural network, arXiv preprint arXiv:1503.02531.
– year: 2019
  ident: b0200
  article-title: Fast human pose estimation
  publication-title: The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
– reference: S. Han, J. Pool, J. Tran, W. Dally, Learning both weights and connections for efficient neural network, in: Proc. Advances in Neural Inf. Process. Syst., 2015, pp. 1135–1143.
– reference: J. Gao, W. Lin, B. Zhao, D. Wang, C. Gao, J. Wen, C-3-Framework: An open-source pytorch code for crowd counting, arXiv preprint arXiv:1907.02724.
– year: 2017
  ident: b0135
  article-title: CNN-based cascaded multi-task learning of high-level prior and density estimation for crowd counting
  publication-title: IEEE International Conference on Advanced Video and Signal Based Surveillance
– year: 2018
  ident: b0235
  article-title: Scale aggregation network for accurate and efficient crowd counting
  publication-title: Proc. Eur. Conf. Comp. Vis.
– year: 2018
  ident: b0080
  article-title: CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 779
  year: 2016
  end-page: 788
  ident: b0125
  article-title: You only look once: unified, real-time object detection
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– reference: V. Nekrasov, H. Chen, C. Shen, I. Reid, Fast neural architecture search of compact semantic segmentation models via auxiliary cells, arXiv preprint arXiv:1810.10804.
– year: 2016
  ident: b0035
  article-title: Efficient piecewise training of deep structured models for semantic segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– reference: P. Viola, M.J. Jones, D. Snow, Detecting pedestrians using patterns of motion and appearance, in: Int. J. Comput. Vision, 2005.
– reference: L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, A.L. Yuille, Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs., IEEE Trans. Pattern Anal. Mach. Intell.
– year: 2018
  ident: b0100
  article-title: Mobilenetv 2: inverted residuals and linear bottlenecks
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2018
  ident: b0085
  article-title: Composition loss for counting, density map estimation and localization in dense crowds
  publication-title: Proc. Eur. Conf. Comp. Vis.
– reference: A.G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, H. Adam, Mobilenets: Efficient convolutional neural networks for mobile vision applications, arXiv preprint arXiv:1704.04861.
– year: 2015
  ident: b0020
  article-title: Very deep convolutional networks for large-scale image recognition
  publication-title: Proc. Int. Conf. Learn. Representations
– year: 2015
  ident: b0030
  article-title: Fully Convolutional Networks for Semantic Segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– reference: F.N. Iandola, S. Han, M.W. Moskewicz, K. Ashraf, W.J. Dally, K. Keutzer, Squeezenet: Alexnet-level accuracy with 50x fewer parameters and< 0.5 mb model size, arXiv preprint arXiv:1602.07360.
– reference: C. Bucilu, R. Caruana, A. Niculescu-Mizil, Model compression, in: Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining, ACM, 2006, pp. 535–541.
– volume: 19
  start-page: 263
  year: 2018
  end-page: 272
  ident: b0110
  article-title: Erfnet: efficient residual factorized convnet for real-time semantic segmentation
  publication-title: IEEE Trans. Intell. Transp. Syst.
– year: 2015
  ident: b0055
  article-title: Cross-scene crowd counting via deep convolutional neural networks
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2016
  ident: b0065
  article-title: Single-image crowd counting via multi-column convolutional neural network
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2016
  ident: b0060
  article-title: Crowdnet: a deep convolutional network for dense crowd counting
  publication-title: Proc. Conf. ACM Multimedia
– year: 2018
  ident: b0075
  article-title: Divide and Grow: Capturing Huge Diversity in Crowd Images with Incrementally Growing, CNN
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2017
  ident: 10.1016/j.neucom.2020.05.056_b0135
  article-title: CNN-based cascaded multi-task learning of high-level prior and density estimation for crowd counting
– start-page: 2117
  year: 2017
  ident: 10.1016/j.neucom.2020.05.056_b0230
  article-title: Feature pyramid networks for object detection
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– ident: 10.1016/j.neucom.2020.05.056_b0105
– ident: 10.1016/j.neucom.2020.05.056_b0130
– ident: 10.1016/j.neucom.2020.05.056_b0155
– ident: 10.1016/j.neucom.2020.05.056_b0180
– ident: 10.1016/j.neucom.2020.05.056_b0015
– ident: 10.1016/j.neucom.2020.05.056_b0040
– start-page: 116
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0160
  article-title: Shufflenet v2: practical guidelines for efficient cnn architecture design
  publication-title: Proc. Eur. Conf. Comp. Vis.
– year: 2009
  ident: 10.1016/j.neucom.2020.05.056_b0050
  article-title: Bayesian poisson regression for crowd counting
  publication-title: Proc. IEEE Int. Conf. Comp. Vis.
– start-page: 4320
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0205
  article-title: Deep mutual learning
– year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0140
  article-title: Crowd counting with deep negative correlation learning
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn
– start-page: 770
  year: 2016
  ident: 10.1016/j.neucom.2020.05.056_b0025
  article-title: Deep residual learning for image recognition
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2016
  ident: 10.1016/j.neucom.2020.05.056_b0060
  article-title: Crowdnet: a deep convolutional network for dense crowd counting
  publication-title: Proc. Conf. ACM Multimedia
– ident: 10.1016/j.neucom.2020.05.056_b0190
– year: 2015
  ident: 10.1016/j.neucom.2020.05.056_b0055
  article-title: Cross-scene crowd counting via deep convolutional neural networks
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 91
  year: 2015
  ident: 10.1016/j.neucom.2020.05.056_b0120
  article-title: Faster r-cnn: Towards real-time object detection with region proposal networks
  publication-title: Adv. Neural Inform. Process. Syst.
– start-page: 6848
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0095
  article-title: Shufflenet: an extremely efficient convolutional neural network for mobile devices
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 3
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0170
  article-title: Convolutional networks with adaptive inference graphs
  publication-title: Proc. Eur. Conf. Comp. Vis.
– year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0235
  article-title: Scale aggregation network for accurate and efficient crowd counting
  publication-title: Proc. Eur. Conf. Comp. Vis.
– year: 2017
  ident: 10.1016/j.neucom.2020.05.056_b0070
  article-title: Switching convolutional neural network for crowd counting
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– ident: 10.1016/j.neucom.2020.05.056_b0195
  doi: 10.1145/1150402.1150464
– start-page: 2654
  year: 2014
  ident: 10.1016/j.neucom.2020.05.056_b0185
  article-title: Do deep nets really need to be deep?
  publication-title: Adv. Neural Inform. Process. Syst.
– year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0085
  article-title: Composition loss for counting, density map estimation and localization in dense crowds
  publication-title: Proc. Eur. Conf. Comp. Vis.
– ident: 10.1016/j.neucom.2020.05.056_b0150
– year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0075
  article-title: Divide and Grow: Capturing Huge Diversity in Crowd Images with Incrementally Growing, CNN
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 6154
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0010
  article-title: Cascade r-cnn: delving into high quality object detection
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2015
  ident: 10.1016/j.neucom.2020.05.056_b0030
  article-title: Fully Convolutional Networks for Semantic Segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0100
  article-title: Mobilenetv 2: inverted residuals and linear bottlenecks
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 7356
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0215
  article-title: Through-wall human pose estimation using radio signals
– year: 2013
  ident: 10.1016/j.neucom.2020.05.056_b0225
  article-title: Multi-source multi-scale counting in extremely dense crowd images
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 779
  year: 2016
  ident: 10.1016/j.neucom.2020.05.056_b0125
  article-title: You only look once: unified, real-time object detection
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 7528
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0210
  article-title: Knowledge distillation by on-the-fly native ensemble
– year: 2016
  ident: 10.1016/j.neucom.2020.05.056_b0035
  article-title: Efficient piecewise training of deep structured models for semantic segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– year: 2015
  ident: 10.1016/j.neucom.2020.05.056_b0020
  article-title: Very deep convolutional networks for large-scale image recognition
  publication-title: Proc. Int. Conf. Learn. Representations
– volume: 19
  start-page: 263
  issue: 1
  year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0110
  article-title: Erfnet: efficient residual factorized convnet for real-time semantic segmentation
  publication-title: IEEE Trans. Intell. Transp. Syst.
  doi: 10.1109/TITS.2017.2750080
– ident: 10.1016/j.neucom.2020.05.056_b0115
– ident: 10.1016/j.neucom.2020.05.056_b0045
  doi: 10.1007/s11263-005-6644-8
– start-page: 580
  year: 2014
  ident: 10.1016/j.neucom.2020.05.056_b0005
  article-title: Rich feature hierarchies for accurate object detection and semantic segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– ident: 10.1016/j.neucom.2020.05.056_b0165
– year: 2018
  ident: 10.1016/j.neucom.2020.05.056_b0080
  article-title: CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– start-page: 1925
  year: 2017
  ident: 10.1016/j.neucom.2020.05.056_b0145
  article-title: Refinenet: multi-path refinement networks for high-resolution semantic segmentation
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– ident: 10.1016/j.neucom.2020.05.056_b0090
– year: 2016
  ident: 10.1016/j.neucom.2020.05.056_b0065
  article-title: Single-image crowd counting via multi-column convolutional neural network
  publication-title: Proc. IEEE Conf. Comp. Vis. Patt. Recogn.
– ident: 10.1016/j.neucom.2020.05.056_b0220
– volume: 18
  start-page: 6869
  issue: 1
  year: 2017
  ident: 10.1016/j.neucom.2020.05.056_b0175
  article-title: Quantized neural networks: training neural networks with low precision weights and activations
  publication-title: J. Mach. Learn. Res.
– year: 2019
  ident: 10.1016/j.neucom.2020.05.056_b0200
  article-title: Fast human pose estimation
  publication-title: The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
SSID ssj0017129
Score 2.5521626
Snippet In this work, we propose a computation-efficient encoder-decoder architecture, named MobileCount, which is specifically designed for high-accuracy real-time...
SourceID crossref
elsevier
SourceType Enrichment Source
Index Database
Publisher
StartPage 292
SubjectTerms Crowd counting
Fully convolutional networks
Knowledge distillation
Light-weight Neural Networks
Title MobileCount: An efficient encoder-decoder framework for real-time crowd counting
URI https://dx.doi.org/10.1016/j.neucom.2020.05.056
Volume 407
WOSCitedRecordID wos000555461000011&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: 1872-8286
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0017129
  issn: 0925-2312
  databaseCode: AIEXJ
  dateStart: 19950101
  isFulltext: true
  titleUrlDefault: https://www.sciencedirect.com
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV1ba9swFBYh3cNetu7GurZDD3srGo4s2VLfwmjZxlYCCyx7MpIls5bULVnSdbAfv6ObY5KxGwyMHYzkCJ3Px5-OzgWhFxxIfGmYJDmVlDBrFNHKFMTImunSqMwY7YtNlGdnYjaTk8Hge4qFuZmXbStub-X1fxU13ANhu9DZvxB391C4Ab9B6HAGscP5jwT__krDm-5izZPVz_o0EW7T32WtNHZBjPXXoya5ZnlvQ-CPc-KKzR-Bdv7qwt1CHYk-gfXJPGpfCiIaGcaXLteCccDqjAofoxV6YmNv5-GjglX2s22_rTYbflLrhu9CIe3Ved8iActPv0mzNpNthcoEeyPlBMhkUL02aFtRUh_H3lfHLFTBTQpV0t63mYZiSltqP1ggLl62duV8gNygfD5WvpFl23-3P7ihuJFQl6tHuhLVO7TkUgzRzvjNyexttwtVjmjI1RiHnkIvvX_g9n_9nNr06Mp0F92L6ww8Dvh4gAa2fYjupxoeOKr0R2jSg8sxHre4AwveAAvuwIIBLLgDC_ZgwQksj9H09GT66jWJVTZIDcvFJYEZEDozvLGN0lxlusgUr4HQNG5tqhpgL5ZxpYBaUyF0bQudZ43kmrm15ih_gobtVWufIiwsFyZnuS6MYFZQKRrajHJWjEwjmLF7KE_zU9UxA70rhDKvkqvhRRVmtXKzWmUcjmIPka7XdcjA8pv2ZZr6KrLIwA4rQMsvez7755776O76RThAw-ViZQ_Rnfpmef5l8TzC6gewPpf8
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=MobileCount%3A+An+efficient+encoder-decoder+framework+for+real-time+crowd+counting&rft.jtitle=Neurocomputing+%28Amsterdam%29&rft.au=Wang%2C+Peng&rft.au=Gao%2C+Chenyu&rft.au=Wang%2C+Yang&rft.au=Li%2C+Hui&rft.date=2020-09-24&rft.pub=Elsevier+B.V&rft.issn=0925-2312&rft.eissn=1872-8286&rft.volume=407&rft.spage=292&rft.epage=299&rft_id=info:doi/10.1016%2Fj.neucom.2020.05.056&rft.externalDocID=S0925231220308912
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0925-2312&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0925-2312&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0925-2312&client=summon