Multi-Task Deep Learning for Real-Time 3D Human Pose Estimation and Action Recognition

Human pose estimation and action recognition are related tasks since both problems are strongly dependent on the human body representation and analysis. Nonetheless, most recent methods in the literature handle the two problems separately. In this article, we propose a multi-task framework for joint...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	IEEE transactions on pattern analysis and machine intelligence Ročník 43; číslo 8; s. 2752 - 2764
Hlavní autoři:	Luvizon, Diogo C., Picard, David, Tabia, Hedi
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	United States IEEE 01.08.2021 The Institute of Electrical and Electronics Engineers, Inc. (IEEE) Institute of Electrical and Electronics Engineers
Témata:	Color imagery Computer Science Decoupling Frames per second Heating systems Human action recognition human pose estimation Image classification Machine learning multitask deep learning neural networks Pose estimation Signal and Image Processing Skeleton Source code Target recognition Task analysis Three-dimensional displays Two dimensional displays Video data Visualization Multitask deep learning Human action recognition Human pose estimation Neural networks
ISSN:	0162-8828, 1939-3539, 2160-9292, 1939-3539
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Abstract	Human pose estimation and action recognition are related tasks since both problems are strongly dependent on the human body representation and analysis. Nonetheless, most recent methods in the literature handle the two problems separately. In this article, we propose a multi-task framework for jointly estimating 2D or 3D human poses from monocular color images and classifying human actions from video sequences. We show that a single architecture can be used to solve both problems in an efficient way and still achieves state-of-the-art or comparable results at each task while running with a throughput of more than 100 frames per second. The proposed method benefits from high parameters sharing between the two tasks by unifying still images and video clips processing in a single pipeline, allowing the model to be trained with data from different categories simultaneously and in a seamlessly way. Additionally, we provide important insights for end-to-end training the proposed multi-task model by decoupling key prediction parts, which consistently leads to better accuracy on both tasks. The reported results on four datasets (MPII, Human3.6M, Penn Action and NTU RGB+D) demonstrate the effectiveness of our method on the targeted tasks. Our source code and trained weights are publicly available at https://github.com/dluvizon/deephar .
AbstractList	Human pose estimation and action recognition are related tasks since both problems are strongly dependent on the human body representation and analysis. Nonetheless, most recent methods in the literature handle the two problems separately. In this work, we propose a multi-task framework for jointly estimating 2D or 3D human poses from monocular color images and classifying human actions from video sequences. We show that a single architecture can be used to solve both problems in an efficient way and still achieves state-of-the-art or comparable results at each task while running with a throughput of more than 100 frames per second. The proposed method benefits from high parameters sharing between the two tasks by unifying still images and video clips processing in a single pipeline, allowing the model to be trained with data from different categories simultaneously and in a seamlessly way. Additionally, we provide important insights for end-to-end training the proposed multi-task model by decoupling key prediction parts, which consistently leads to better accuracy on both tasks. The reported results on four datasets (MPII, Human3.6M, Penn Action and NTU RGB+D) demonstrate the effectiveness of our method on the targeted tasks. Our source code and trained weights are publicly available at https://github.com/dluvizon/deephar. Human pose estimation and action recognition are related tasks since both problems are strongly dependent on the human body representation and analysis. Nonetheless, most recent methods in the literature handle the two problems separately. In this article, we propose a multi-task framework for jointly estimating 2D or 3D human poses from monocular color images and classifying human actions from video sequences. We show that a single architecture can be used to solve both problems in an efficient way and still achieves state-of-the-art or comparable results at each task while running with a throughput of more than 100 frames per second. The proposed method benefits from high parameters sharing between the two tasks by unifying still images and video clips processing in a single pipeline, allowing the model to be trained with data from different categories simultaneously and in a seamlessly way. Additionally, we provide important insights for end-to-end training the proposed multi-task model by decoupling key prediction parts, which consistently leads to better accuracy on both tasks. The reported results on four datasets (MPII, Human3.6M, Penn Action and NTU RGB+D) demonstrate the effectiveness of our method on the targeted tasks. Our source code and trained weights are publicly available at https://github.com/dluvizon/deephar . Human pose estimation and action recognition are related tasks since both problems are strongly dependent on the human body representation and analysis. Nonetheless, most recent methods in the literature handle the two problems separately. In this article, we propose a multi-task framework for jointly estimating 2D or 3D human poses from monocular color images and classifying human actions from video sequences. We show that a single architecture can be used to solve both problems in an efficient way and still achieves state-of-the-art or comparable results at each task while running with a throughput of more than 100 frames per second. The proposed method benefits from high parameters sharing between the two tasks by unifying still images and video clips processing in a single pipeline, allowing the model to be trained with data from different categories simultaneously and in a seamlessly way. Additionally, we provide important insights for end-to-end training the proposed multi-task model by decoupling key prediction parts, which consistently leads to better accuracy on both tasks. The reported results on four datasets (MPII, Human3.6M, Penn Action and NTU RGB+D) demonstrate the effectiveness of our method on the targeted tasks. Our source code and trained weights are publicly available at https://github.com/dluvizon/deephar.Human pose estimation and action recognition are related tasks since both problems are strongly dependent on the human body representation and analysis. Nonetheless, most recent methods in the literature handle the two problems separately. In this article, we propose a multi-task framework for jointly estimating 2D or 3D human poses from monocular color images and classifying human actions from video sequences. We show that a single architecture can be used to solve both problems in an efficient way and still achieves state-of-the-art or comparable results at each task while running with a throughput of more than 100 frames per second. The proposed method benefits from high parameters sharing between the two tasks by unifying still images and video clips processing in a single pipeline, allowing the model to be trained with data from different categories simultaneously and in a seamlessly way. Additionally, we provide important insights for end-to-end training the proposed multi-task model by decoupling key prediction parts, which consistently leads to better accuracy on both tasks. The reported results on four datasets (MPII, Human3.6M, Penn Action and NTU RGB+D) demonstrate the effectiveness of our method on the targeted tasks. Our source code and trained weights are publicly available at https://github.com/dluvizon/deephar.
Author	Tabia, Hedi Luvizon, Diogo C. Picard, David
Author_xml	– sequence: 1 givenname: Diogo C. orcidid: 0000-0002-5055-500X surname: Luvizon fullname: Luvizon, Diogo C. email: diogo.luvizon@ensea.fr organization: SAMSUNG Research Institute, Campinas, SP, Brazil – sequence: 2 givenname: David orcidid: 0000-0002-6296-4222 surname: Picard fullname: Picard, David email: david.picard@enpc.fr organization: LIGM, IMAGINE, Ecole des Ponts, Univ Gustave Eiffel, CNRS, Marne-la-Vallée, France – sequence: 3 givenname: Hedi surname: Tabia fullname: Tabia, Hedi email: hedi.tabia@univ-evry.fr organization: IBISC, Univ Evry, Université Paris-Saclay, Evry, France
BackLink	https://www.ncbi.nlm.nih.gov/pubmed/32091993$$D View this record in MEDLINE/PubMed https://hal.science/hal-02558843$$DView record in HAL
BookMark	eNp9kcFu1DAQhi1URLeFFwAJWeIChyz2OHHs46otbKWtqKptr5bXmRSXxF7ipFLfnmR320MPnDyyvm88nv-EHIUYkJCPnM05Z_r7-npxdTkHBmwOupSM52_IDLhkmQYNR2TGuIRMKVDH5CSlBzYSBRPvyLEAprnWYkburoam99napj_0HHFLV2i74MM9rWNHb9A22dq3SMU5XQ6tDfQ6JqQXqfet7X0M1IaKLtyuvEEX74Of6vfkbW2bhB8O5ym5_XGxPltmq18_L88Wq8wJVfQZlFUldc0U2FK7QjqXQyG1kgyklXVhkWspWb1hFbiNlhvtVF6jBp7XFdhanJJv-76_bWO23ThU92Si9Wa5WJnpjkFRKJWLRz6yX_fstot_B0y9aX1y2DQ2YBySASFzJhTXE_rlFfoQhy6MPzFQ5KXM81LBSH0-UMOmxerl_eftjoDaA66LKXVYG-f73dr6zvrGcGamIM0uSDMFaQ5Bjiq8Up-7_1f6tJc8Ir4ImrFS6kL8A2CWpXQ
CODEN	ITPIDJ
CitedBy_id	crossref_primary_10_1109_JSEN_2023_3315849 crossref_primary_10_1109_ACCESS_2022_3186465 crossref_primary_10_1109_TASE_2023_3279928 crossref_primary_10_1109_TPAMI_2023_3238411 crossref_primary_10_1007_s10489_022_03623_z crossref_primary_10_1016_j_image_2021_116410 crossref_primary_10_1109_TIM_2021_3092524 crossref_primary_10_1016_j_neucom_2025_131309 crossref_primary_10_3389_fpsyt_2022_1019043 crossref_primary_10_1109_ACCESS_2023_3241606 crossref_primary_10_1016_j_eswa_2021_116424 crossref_primary_10_1109_TPAMI_2024_3507918 crossref_primary_10_1007_s00138_020_01120_2 crossref_primary_10_1109_TNNLS_2022_3175480 crossref_primary_10_4018_IJAEC_315633 crossref_primary_10_1109_JSEN_2023_3323869 crossref_primary_10_1109_ACCESS_2024_3470789 crossref_primary_10_1049_ipr2_12404 crossref_primary_10_1109_TCSVT_2023_3284493 crossref_primary_10_3389_fnbot_2024_1371385 crossref_primary_10_20965_jaciii_2024_p0552 crossref_primary_10_1109_TPAMI_2024_3364185 crossref_primary_10_1016_j_patcog_2021_108487 crossref_primary_10_1109_TNNLS_2023_3264647 crossref_primary_10_1109_TPAMI_2022_3188716 crossref_primary_10_1109_TMM_2024_3521749 crossref_primary_10_1155_2021_5593916 crossref_primary_10_1515_jisys_2024_0082 crossref_primary_10_1109_TPAMI_2022_3170353 crossref_primary_10_1109_TIE_2021_3105977 crossref_primary_10_1007_s13369_022_07236_z crossref_primary_10_1007_s11042_020_09708_6 crossref_primary_10_1007_s11263_021_01529_w
Cites_doi	10.1109/ICCV.2013.396 10.1007/978-3-319-46478-7_44 10.1109/CVPR.2018.00539 10.1109/CVPR.2017.391 10.1109/CVPR.2018.00127 10.1109/ICCV.2013.280 10.1109/CVPR.2018.00056 10.1109/CVPR.2016.115 10.1109/TPAMI.2017.2712608 10.5244/C.30.109 10.1109/ICCV.2017.137 10.1109/CVPR.2017.195 10.1109/3DV.2017.00064 10.1109/CVPR.2013.391 10.1109/CVPR.2017.601 10.1109/TPAMI.2013.248 10.1109/CVPR.2017.610 10.1016/j.cviu.2016.09.002 10.1109/TPAMI.2018.2816031 10.1109/FG.2017.61 10.1109/CVPR.2017.139 10.1016/j.patcog.2015.11.019 10.1109/CVPR.2016.511 10.1109/CVPR.2015.7298664 10.1016/j.patrec.2017.02.001 10.1109/CVPR.2016.533 10.1109/ICCV.2017.284 10.1109/CVPR.2016.512 10.1016/j.imavis.2017.01.010 10.1016/j.cag.2019.09.002 10.1109/CVPR.2017.603 10.1109/CVPR.2013.82 10.1109/CVPR.2019.00584 10.1007/978-3-030-01231-1_33 10.1109/CVPR.2009.5206754 10.1109/CVPR.2017.486 10.1007/s11263-012-0532-9 10.1007/978-3-319-46466-4_28 10.1007/978-3-030-01240-3_28 10.1109/CVPR.2017.579 10.1007/978-3-319-46493-0_44 10.1109/ICCV.2017.144 10.1109/CVPR.2017.501 10.1109/CVPR.2014.471 10.1109/TCYB.2017.2756840 10.1109/ICCV.2017.288 10.1109/TPAMI.2017.2691321 10.1145/3072959.3073596 10.1109/CVPR.2018.00551 10.1007/978-3-319-46466-4_3 10.1007/978-3-030-01252-6_8 10.1109/ICCV.2015.324 10.1109/CVPR.2014.214 10.1109/TMM.2017.2762010 10.1111/j.1467-9868.2005.00503.x 10.1109/ICCV.2017.402 10.1109/ICCV.2017.316 10.1109/CVPR.2018.00734 10.1109/CVPR.2017.502
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2021 Distributed under a Creative Commons Attribution 4.0 International License
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2021 – notice: Distributed under a Creative Commons Attribution 4.0 International License
DBID	97E RIA RIE AAYXX CITATION NPM 7SC 7SP 8FD JQ2 L7M L~C L~D 7X8 1XC VOOES
DOI	10.1109/TPAMI.2020.2976014
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef PubMed Computer and Information Systems Abstracts Electronics & Communications Abstracts Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional MEDLINE - Academic Hyper Article en Ligne (HAL) Hyper Article en Ligne (HAL) (Open Access)
DatabaseTitle	CrossRef PubMed Technology Research Database Computer and Information Systems Abstracts – Academic Electronics & Communications Abstracts ProQuest Computer Science Collection Computer and Information Systems Abstracts Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Professional MEDLINE - Academic
DatabaseTitleList	Technology Research Database MEDLINE - Academic PubMed
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: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering Computer Science
EISSN	2160-9292 1939-3539
EndPage	2764
ExternalDocumentID	oai:HAL:hal-02558843v1 32091993 10_1109_TPAMI_2020_2976014 9007695
Genre	orig-research Journal Article
GrantInformation_xml	– fundername: Brazilian National Council for Scientific and Technological Development grantid: 233342/2014-1
GroupedDBID	--- -DZ -~X .DC 0R~ 29I 4.4 53G 5GY 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACGFO ACGFS ACIWK ACNCT AENEX AGQYO AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 E.L EBS EJD F5P HZ~ IEDLZ IFIPE IPLJI JAVBF LAI M43 MS~ O9- OCL P2P PQQKQ RIA RIE RNS RXW TAE TN5 UHB ~02 5VS 9M8 AAYXX ABFSI ADRHT AETEA AETIX AGSQL AI. AIBXA ALLEH CITATION FA8 H~9 IBMZZ ICLAB IFJZH RNI RZB VH1 NPM 7SC 7SP 8FD JQ2 L7M L~C L~D 7X8 1XC VOOES XJT
ID	FETCH-LOGICAL-c385t-27dd69f082a79c56cc4256986026a6f5ae19660fb0d2cb96b9c84fe9214fd2af3
IEDL.DBID	RIE
ISICitedReferencesCount	109
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000670578800017&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	0162-8828 1939-3539
IngestDate	Sun Oct 19 06:20:44 EDT 2025 Sun Sep 28 06:30:53 EDT 2025 Sun Jun 29 12:31:43 EDT 2025 Mon Jul 21 06:03:21 EDT 2025 Sat Nov 29 05:15:59 EST 2025 Tue Nov 18 22:34:45 EST 2025 Wed Aug 27 02:26:41 EDT 2025
IsDoiOpenAccess	true
IsOpenAccess	true
IsPeerReviewed	true
IsScholarly	true
Issue	8
Keywords	Multitask deep learning Human action recognition Human pose estimation Neural networks
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 Distributed under a Creative Commons Attribution 4.0 International License: http://creativecommons.org/licenses/by/4.0
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c385t-27dd69f082a79c56cc4256986026a6f5ae19660fb0d2cb96b9c84fe9214fd2af3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 content type line 23
ORCID	0000-0002-6296-4222 0000-0002-5055-500X 0000-0002-1827-7150
OpenAccessLink	https://hal.science/hal-02558843
PMID	32091993
PQID	2547644782
PQPubID	85458
PageCount	13
ParticipantIDs	pubmed_primary_32091993 crossref_citationtrail_10_1109_TPAMI_2020_2976014 proquest_journals_2547644782 crossref_primary_10_1109_TPAMI_2020_2976014 proquest_miscellaneous_2364038191 ieee_primary_9007695 hal_primary_oai_HAL_hal_02558843v1
PublicationCentury	2000
PublicationDate	2021-08-01
PublicationDateYYYYMMDD	2021-08-01
PublicationDate_xml	– month: 08 year: 2021 text: 2021-08-01 day: 01
PublicationDecade	2020
PublicationPlace	United States
PublicationPlace_xml	– name: United States – name: New York
PublicationTitle	IEEE transactions on pattern analysis and machine intelligence
PublicationTitleAbbrev	TPAMI
PublicationTitleAlternate	IEEE Trans Pattern Anal Mach Intell
PublicationYear	2021
Publisher	IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE) Institute of Electrical and Electronics Engineers
Publisher_xml	– name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE) – name: Institute of Electrical and Electronics Engineers
References	ref57 ref13 ref56 ref12 ref59 ref15 ref58 ref14 ref53 ref52 ref55 ref11 baradel (ref64) 2017 ref54 ref10 ref17 ref19 ref18 ref51 ref50 ref46 ref45 ref48 ref47 ref42 tekin (ref38) 2016; abs 1611 5708 ref41 ref44 ref43 ref8 ref7 ref9 ref4 chou (ref32) 2017 ref3 ref6 chéron (ref1) 2015 ref5 ref40 liu (ref60) 2016 ref35 ref34 ref37 lifshitz (ref16) 2016 ref36 ref30 ref33 newell (ref27) 2016 ref2 ref39 ref71 ref70 song (ref62) 2017 ref68 ref67 ref23 ref26 ref69 ref25 ref20 ref63 ref66 ref22 ref65 ref21 goodfellow (ref31) 2014 ref28 pfister (ref24) 2014 ref29 xiaohan nie (ref49) 2015 ref61
References_xml	– start-page: 3218 year: 2015 ident: ref1 article-title: P-CNN: Pose-based CNN features for action recognition publication-title: Proc ICCV – ident: ref50 doi: 10.1109/ICCV.2013.396 – ident: ref25 doi: 10.1007/978-3-319-46478-7_44 – ident: ref9 doi: 10.1109/CVPR.2018.00539 – ident: ref61 doi: 10.1109/CVPR.2017.391 – ident: ref57 doi: 10.1109/CVPR.2018.00127 – ident: ref68 doi: 10.1109/ICCV.2013.280 – ident: ref55 doi: 10.1109/CVPR.2018.00056 – ident: ref69 doi: 10.1109/CVPR.2016.115 – ident: ref53 doi: 10.1109/TPAMI.2017.2712608 – ident: ref19 doi: 10.5244/C.30.109 – ident: ref33 doi: 10.1109/ICCV.2017.137 – start-page: 816 year: 2016 ident: ref60 article-title: Spatio-temporal LSTM with trust gates for 3D human action recognition publication-title: Proc Eur Conf Comput Vis – ident: ref65 doi: 10.1109/CVPR.2017.195 – start-page: 538 year: 2014 ident: ref24 article-title: Deep convolutional neural networks for efficient pose estimation in gesture videos publication-title: Proc Asian Conf Comput Vis – ident: ref39 doi: 10.1109/3DV.2017.00064 – ident: ref13 doi: 10.1109/CVPR.2013.391 – ident: ref28 doi: 10.1109/CVPR.2017.601 – year: 2017 ident: ref64 article-title: Pose-conditioned spatio-temporal attention for human action recognition – ident: ref41 doi: 10.1109/TPAMI.2013.248 – ident: ref43 doi: 10.1109/CVPR.2017.610 – ident: ref71 doi: 10.1109/CVPR.2018.00056 – start-page: 483 year: 2016 ident: ref27 article-title: Stacked hourglass networks for human pose estimation publication-title: Proc Eur Conf Comput Vis – ident: ref10 doi: 10.1016/j.cviu.2016.09.002 – ident: ref35 doi: 10.1109/TPAMI.2018.2816031 – ident: ref3 doi: 10.1109/FG.2017.61 – ident: ref45 doi: 10.1109/CVPR.2017.139 – ident: ref59 doi: 10.1016/j.patcog.2015.11.019 – start-page: 4263 year: 2017 ident: ref62 article-title: An end-to-end spatio-temporal attention model for human action recognition from skeleton data publication-title: Proc 31st AAAI Conf Artif Intell – ident: ref20 doi: 10.1109/CVPR.2016.511 – ident: ref22 doi: 10.1109/CVPR.2015.7298664 – ident: ref58 doi: 10.1016/j.patrec.2017.02.001 – ident: ref17 doi: 10.1109/CVPR.2016.533 – ident: ref44 doi: 10.1109/ICCV.2017.284 – ident: ref34 doi: 10.1109/CVPR.2016.512 – ident: ref11 doi: 10.1016/j.imavis.2017.01.010 – ident: ref7 doi: 10.1016/j.cag.2019.09.002 – volume: abs 1611 5708 year: 2016 ident: ref38 article-title: Fusing 2D uncertainty and 3D cues for monocular body pose estimation publication-title: CoRR – ident: ref36 doi: 10.1109/CVPR.2017.603 – ident: ref14 doi: 10.1109/CVPR.2013.82 – ident: ref30 doi: 10.1109/CVPR.2019.00584 – ident: ref46 doi: 10.1007/978-3-030-01231-1_33 – ident: ref12 doi: 10.1109/CVPR.2009.5206754 – ident: ref66 doi: 10.1109/CVPR.2017.486 – ident: ref2 doi: 10.1007/s11263-012-0532-9 – ident: ref8 doi: 10.1007/978-3-319-46466-4_28 – ident: ref56 doi: 10.1007/978-3-030-01240-3_28 – ident: ref4 doi: 10.1109/CVPR.2017.579 – ident: ref26 doi: 10.1007/978-3-319-46493-0_44 – ident: ref29 doi: 10.1109/ICCV.2017.144 – ident: ref40 doi: 10.1109/CVPR.2017.501 – start-page: 246 year: 2016 ident: ref16 publication-title: Human Pose Estimation using Deep Consensus Voting – ident: ref67 doi: 10.1109/CVPR.2014.471 – ident: ref51 doi: 10.1109/TCYB.2017.2756840 – ident: ref37 doi: 10.1109/ICCV.2017.288 – ident: ref63 doi: 10.1109/TPAMI.2017.2691321 – ident: ref42 doi: 10.1145/3072959.3073596 – start-page: 1293 year: 2015 ident: ref49 article-title: Joint action recognition and pose estimation from video publication-title: Proc IEEE Conf Comput Vis Pattern Recognit – ident: ref47 doi: 10.1109/CVPR.2018.00551 – ident: ref18 doi: 10.1007/978-3-319-46466-4_3 – ident: ref48 doi: 10.1007/978-3-030-01252-6_8 – start-page: 2672 year: 2014 ident: ref31 article-title: Generative adversarial nets publication-title: Proc Int Conf Neural Inf Process – ident: ref21 doi: 10.1109/ICCV.2015.324 – ident: ref23 doi: 10.1109/CVPR.2014.214 – ident: ref15 doi: 10.1109/TMM.2017.2762010 – ident: ref70 doi: 10.1111/j.1467-9868.2005.00503.x – ident: ref54 doi: 10.1109/ICCV.2017.402 – start-page: 17 year: 2017 ident: ref32 article-title: Self adversarial training for human pose estimation publication-title: Proc Asia-Pacific Signal Inf Process Assoc Annu Summit Conf – ident: ref5 doi: 10.1109/ICCV.2017.316 – ident: ref6 doi: 10.1109/CVPR.2018.00734 – ident: ref52 doi: 10.1109/CVPR.2017.502
SSID	ssj0014503
Score	2.6745834
Snippet	Human pose estimation and action recognition are related tasks since both problems are strongly dependent on the human body representation and analysis....
SourceID	hal proquest pubmed crossref ieee
SourceType	Open Access Repository Aggregation Database Index Database Enrichment Source Publisher
StartPage	2752
SubjectTerms	Color imagery Computer Science Decoupling Frames per second Heating systems Human action recognition human pose estimation Image classification Machine learning multitask deep learning neural networks Pose estimation Signal and Image Processing Skeleton Source code Target recognition Task analysis Three-dimensional displays Two dimensional displays Video data Visualization
Title	Multi-Task Deep Learning for Real-Time 3D Human Pose Estimation and Action Recognition
URI	https://ieeexplore.ieee.org/document/9007695 https://www.ncbi.nlm.nih.gov/pubmed/32091993 https://www.proquest.com/docview/2547644782 https://www.proquest.com/docview/2364038191 https://hal.science/hal-02558843
Volume	43
WOSCitedRecordID	wos000670578800017&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: 2160-9292 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0014503 issn: 0162-8828 databaseCode: RIE dateStart: 19790101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT9wwEB6xiEN7KC30kRaQW_XWGhLbcezjqoBAomiFtmhvke04bQVKELvL7-_YeaiVWqTeosROosxM_H2eF8BHq3nFMquotCalwhhBlXWMSuGVKaTIbSzXdH1RXF6qxULPNuDzmAvjvY_BZ_4wHEZfftW6ddgqO9LBb6TzCUyKouhytUaPgchjF2REMGjhSCOGBJlUH81n06_nSAVZesh0iAEJzXg4w5VSa_7HejT5EaIhY5uVfyPOuPKcbv_fOz-HZz3CJNNOJV7Ahm92YHvo3kB6Y96Bp7-VItyF65iJS-dmeUOOvb8jfeHV7wRRLblCOElDtgjhxyTu-5NZu_TkBH8QXe4jMU1FpjFJglwNQUlt8xK-nZ7Mv5zRvucCdVzlK8qKqpK6RmBgCu1y6RwatdSxU5WRdW58Fup51jatmLNaWu2UqL1mmagrZmr-CjabtvFvgGRauVRYy5GECcmdrWuJ6I9xZDA4gSWQDV--dH1B8tAX47aMxCTVZRRcGQRX9oJL4NM4564rx_Ho6A8o0HFgqKR9Nr0ow7lApZQS_CFLYDdIbRzVCyyBvUH-ZW_WyxLZNOqvQFSVwPvxMhpk8LKYxrdrHMOlSCMPTuB1pzfjvQele_v3Z76DJyyEzMT4wj3YXN2v_T5suYfVz-X9AWr9Qh1Erf8FJEv3Sg
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Lb9QwEB61BYlyoNDyCBQwiBtNm9iO1z6uaKut2K5W1VL1ZtmOAxUoqbq7_f2MnYdAAiRuUWInUWYm_j7PC-CDVaykuZWpsCZLuTE8ldbRVHAvzUjwwsZyTZfT0Wwmr67UfAMOhlwY730MPvOH4TD68svGrcNW2ZEKfiNVbMK9gnOat9lag8-AF7EPMmIYtHEkEn2KTKaOFvPx-RmSQZodUhWiQEI7HkZxrVSK_bYibX4L8ZCx0crfMWdce053_u-tH8OjDmOScasUT2DD17uw0_dvIJ0578LDX4oR7sFlzMVNF2b5nRx7f0O60qtfCeJacoGAMg35IoQdk7jzT-bN0pMT_EW02Y_E1CUZxzQJctGHJTX1U_hyerL4NEm7rgupY7JYpXRUlkJVCA3MSLlCOIdmLVTsVWVEVRifh4qelc1K6qwSVjnJK69ozquSmoo9g626qf0LILmSLuPWMqRhXDBnq0og_qMMOQxOoAnk_ZfXritJHjpj_NCRmmRKR8HpIDjdCS6Bj8Ocm7Ygxz9Hv0eBDgNDLe3JeKrDuUCmpOTsLk9gL0htGNUJLIH9Xv66M-ylRj6NGswRVyXwbriMJhn8LKb2zRrHMMGzyIQTeN7qzXDvXule_vmZb-HBZHE-1dOz2edXsE1DAE2MNtyHrdXt2r-G--5udb28fRN1_ydUm_mp
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=Multi-Task+Deep+Learning+for+Real-Time+3D+Human+Pose+Estimation+and+Action+Recognition&rft.jtitle=IEEE+transactions+on+pattern+analysis+and+machine+intelligence&rft.au=Luvizon%2C+Diogo+C&rft.au=Picard%2C+David&rft.au=Tabia%2C+Hedi&rft.date=2021-08-01&rft.pub=The+Institute+of+Electrical+and+Electronics+Engineers%2C+Inc.+%28IEEE%29&rft.issn=0162-8828&rft.eissn=1939-3539&rft.volume=43&rft.issue=8&rft.spage=2752&rft_id=info:doi/10.1109%2FTPAMI.2020.2976014&rft.externalDBID=NO_FULL_TEXT
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0162-8828&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0162-8828&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0162-8828&client=summon