Stacked hourglass networks for human pose estimation

Alejandro Newell; Kaiyu Yang; Jia Deng

doi:10.1007/978-3-319-46484-8_29

Stacked hourglass networks for human pose estimation

Alejandro Newell, Kaiyu Yang, Jia Deng

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

3638 Scopus citations

Abstract

This work introduces a novel convolutional network architecture for the task of human pose estimation. Features are processed across all scales and consolidated to best capture the various spatial relationships associated with the body. We show how repeated bottom-up, top-down processing used in conjunction with intermediate supervision is critical to improving the performance of the network. We refer to the architecture as a “stacked hourglass” network based on the successive steps of pooling and upsampling that are done to produce a final set of predictions. State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods.

Original language	English (US)
Title of host publication	Computer Vision - 14th European Conference, ECCV 2016, Proceedings
Editors	Bastian Leibe, Jiri Matas, Nicu Sebe, Max Welling
Publisher	Springer Verlag
Pages	483-499
Number of pages	17
ISBN (Print)	9783319464831
DOIs	https://doi.org/10.1007/978-3-319-46484-8_29
State	Published - 2016
Externally published	Yes
Event	14th European Conference on Computer Vision, ECCV 2016 - Amsterdam, Netherlands Duration: Oct 8 2016 → Oct 16 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9912 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th European Conference on Computer Vision, ECCV 2016
Country/Territory	Netherlands
City	Amsterdam
Period	10/8/16 → 10/16/16

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Keywords

Human pose estimation

Access to Document

10.1007/978-3-319-46484-8_29

Cite this

Newell, A., Yang, K., & Deng, J. (2016). Stacked hourglass networks for human pose estimation. In B. Leibe, J. Matas, N. Sebe, & M. Welling (Eds.), Computer Vision - 14th European Conference, ECCV 2016, Proceedings (pp. 483-499). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9912 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-46484-8_29

Newell, Alejandro ; Yang, Kaiyu ; Deng, Jia. / Stacked hourglass networks for human pose estimation. Computer Vision - 14th European Conference, ECCV 2016, Proceedings. editor / Bastian Leibe ; Jiri Matas ; Nicu Sebe ; Max Welling. Springer Verlag, 2016. pp. 483-499 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{870a37a4f4e24a66bce3afcddaadf871,

title = "Stacked hourglass networks for human pose estimation",

abstract = "This work introduces a novel convolutional network architecture for the task of human pose estimation. Features are processed across all scales and consolidated to best capture the various spatial relationships associated with the body. We show how repeated bottom-up, top-down processing used in conjunction with intermediate supervision is critical to improving the performance of the network. We refer to the architecture as a “stacked hourglass” network based on the successive steps of pooling and upsampling that are done to produce a final set of predictions. State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods.",

keywords = "Human pose estimation",

author = "Alejandro Newell and Kaiyu Yang and Jia Deng",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2016.; 14th European Conference on Computer Vision, ECCV 2016 ; Conference date: 08-10-2016 Through 16-10-2016",

year = "2016",

doi = "10.1007/978-3-319-46484-8_29",

language = "English (US)",

isbn = "9783319464831",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "483--499",

editor = "Bastian Leibe and Jiri Matas and Nicu Sebe and Max Welling",

booktitle = "Computer Vision - 14th European Conference, ECCV 2016, Proceedings",

address = "Germany",

}

Newell, A, Yang, K & Deng, J 2016, Stacked hourglass networks for human pose estimation. in B Leibe, J Matas, N Sebe & M Welling (eds), Computer Vision - 14th European Conference, ECCV 2016, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9912 LNCS, Springer Verlag, pp. 483-499, 14th European Conference on Computer Vision, ECCV 2016, Amsterdam, Netherlands, 10/8/16. https://doi.org/10.1007/978-3-319-46484-8_29

Stacked hourglass networks for human pose estimation. / Newell, Alejandro; Yang, Kaiyu; Deng, Jia.
Computer Vision - 14th European Conference, ECCV 2016, Proceedings. ed. / Bastian Leibe; Jiri Matas; Nicu Sebe; Max Welling. Springer Verlag, 2016. p. 483-499 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9912 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Stacked hourglass networks for human pose estimation

AU - Newell, Alejandro

AU - Yang, Kaiyu

AU - Deng, Jia

N1 - Publisher Copyright: © Springer International Publishing AG 2016.

PY - 2016

Y1 - 2016

N2 - This work introduces a novel convolutional network architecture for the task of human pose estimation. Features are processed across all scales and consolidated to best capture the various spatial relationships associated with the body. We show how repeated bottom-up, top-down processing used in conjunction with intermediate supervision is critical to improving the performance of the network. We refer to the architecture as a “stacked hourglass” network based on the successive steps of pooling and upsampling that are done to produce a final set of predictions. State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods.

AB - This work introduces a novel convolutional network architecture for the task of human pose estimation. Features are processed across all scales and consolidated to best capture the various spatial relationships associated with the body. We show how repeated bottom-up, top-down processing used in conjunction with intermediate supervision is critical to improving the performance of the network. We refer to the architecture as a “stacked hourglass” network based on the successive steps of pooling and upsampling that are done to produce a final set of predictions. State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods.

KW - Human pose estimation

UR - http://www.scopus.com/inward/record.url?scp=84990052811&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84990052811&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-46484-8_29

DO - 10.1007/978-3-319-46484-8_29

M3 - Conference contribution

AN - SCOPUS:84990052811

SN - 9783319464831

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 483

EP - 499

BT - Computer Vision - 14th European Conference, ECCV 2016, Proceedings

A2 - Leibe, Bastian

A2 - Matas, Jiri

A2 - Sebe, Nicu

A2 - Welling, Max

PB - Springer Verlag

T2 - 14th European Conference on Computer Vision, ECCV 2016

Y2 - 8 October 2016 through 16 October 2016

ER -

Newell A, Yang K, Deng J. Stacked hourglass networks for human pose estimation. In Leibe B, Matas J, Sebe N, Welling M, editors, Computer Vision - 14th European Conference, ECCV 2016, Proceedings. Springer Verlag. 2016. p. 483-499. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-46484-8_29

Stacked hourglass networks for human pose estimation

Abstract

Publication series

Conference

All Science Journal Classification (ASJC) codes

Keywords

Access to Document

Other files and links

Fingerprint

Cite this