A Closer Look at Spatiotemporal Convolutions for Action Recognition

Du Tran, Heng Wang, Lorenzo Torresani, Jamie Ray, Yann LeCun, Manohar Paluri

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper we discuss several forms of spatiotemporal convolutions for video analysis and study their effects on action recognition. Our motivation stems from the observation that 2D CNNs applied to individual frames of the video have remained solid performers in action recognition. In this work we empirically demonstrate the accuracy advantages of 3D CNNs over 2D CNNs within the framework of residual learning. Furthermore, we show that factorizing the 3D convolutional filters into separate spatial and temporal components yields significantly gains in accuracy. Our empirical study leads to the design of a new spatiotemporal convolutional block 'R(2+1)D' which produces CNNs that achieve results comparable or superior to the state-of-the-art on Sports-1M, Kinetics, UCF101, and HMDB51.

Original languageEnglish (US)
Title of host publicationProceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018
PublisherIEEE Computer Society
Pages6450-6459
Number of pages10
ISBN (Electronic)9781538664209
DOIs
StatePublished - Dec 14 2018
Event31st Meeting of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018 - Salt Lake City, United States
Duration: Jun 18 2018Jun 22 2018

Publication series

NameProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
ISSN (Print)1063-6919

Conference

Conference31st Meeting of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018
CountryUnited States
CitySalt Lake City
Period6/18/186/22/18

Fingerprint

Sports
Convolution
Kinetics

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this

Tran, D., Wang, H., Torresani, L., Ray, J., LeCun, Y., & Paluri, M. (2018). A Closer Look at Spatiotemporal Convolutions for Action Recognition. In Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018 (pp. 6450-6459). [8578773] (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition). IEEE Computer Society. https://doi.org/10.1109/CVPR.2018.00675

A Closer Look at Spatiotemporal Convolutions for Action Recognition. / Tran, Du; Wang, Heng; Torresani, Lorenzo; Ray, Jamie; LeCun, Yann; Paluri, Manohar.

Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018. IEEE Computer Society, 2018. p. 6450-6459 8578773 (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Tran, D, Wang, H, Torresani, L, Ray, J, LeCun, Y & Paluri, M 2018, A Closer Look at Spatiotemporal Convolutions for Action Recognition. in Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018., 8578773, Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, IEEE Computer Society, pp. 6450-6459, 31st Meeting of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018, Salt Lake City, United States, 6/18/18. https://doi.org/10.1109/CVPR.2018.00675
Tran D, Wang H, Torresani L, Ray J, LeCun Y, Paluri M. A Closer Look at Spatiotemporal Convolutions for Action Recognition. In Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018. IEEE Computer Society. 2018. p. 6450-6459. 8578773. (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition). https://doi.org/10.1109/CVPR.2018.00675
Tran, Du ; Wang, Heng ; Torresani, Lorenzo ; Ray, Jamie ; LeCun, Yann ; Paluri, Manohar. / A Closer Look at Spatiotemporal Convolutions for Action Recognition. Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018. IEEE Computer Society, 2018. pp. 6450-6459 (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition).
@inproceedings{333077d3a79c4a098f9c777aa1881a8a,
title = "A Closer Look at Spatiotemporal Convolutions for Action Recognition",
abstract = "In this paper we discuss several forms of spatiotemporal convolutions for video analysis and study their effects on action recognition. Our motivation stems from the observation that 2D CNNs applied to individual frames of the video have remained solid performers in action recognition. In this work we empirically demonstrate the accuracy advantages of 3D CNNs over 2D CNNs within the framework of residual learning. Furthermore, we show that factorizing the 3D convolutional filters into separate spatial and temporal components yields significantly gains in accuracy. Our empirical study leads to the design of a new spatiotemporal convolutional block 'R(2+1)D' which produces CNNs that achieve results comparable or superior to the state-of-the-art on Sports-1M, Kinetics, UCF101, and HMDB51.",
author = "Du Tran and Heng Wang and Lorenzo Torresani and Jamie Ray and Yann LeCun and Manohar Paluri",
year = "2018",
month = "12",
day = "14",
doi = "10.1109/CVPR.2018.00675",
language = "English (US)",
series = "Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
publisher = "IEEE Computer Society",
pages = "6450--6459",
booktitle = "Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018",

}

TY - GEN

T1 - A Closer Look at Spatiotemporal Convolutions for Action Recognition

AU - Tran, Du

AU - Wang, Heng

AU - Torresani, Lorenzo

AU - Ray, Jamie

AU - LeCun, Yann

AU - Paluri, Manohar

PY - 2018/12/14

Y1 - 2018/12/14

N2 - In this paper we discuss several forms of spatiotemporal convolutions for video analysis and study their effects on action recognition. Our motivation stems from the observation that 2D CNNs applied to individual frames of the video have remained solid performers in action recognition. In this work we empirically demonstrate the accuracy advantages of 3D CNNs over 2D CNNs within the framework of residual learning. Furthermore, we show that factorizing the 3D convolutional filters into separate spatial and temporal components yields significantly gains in accuracy. Our empirical study leads to the design of a new spatiotemporal convolutional block 'R(2+1)D' which produces CNNs that achieve results comparable or superior to the state-of-the-art on Sports-1M, Kinetics, UCF101, and HMDB51.

AB - In this paper we discuss several forms of spatiotemporal convolutions for video analysis and study their effects on action recognition. Our motivation stems from the observation that 2D CNNs applied to individual frames of the video have remained solid performers in action recognition. In this work we empirically demonstrate the accuracy advantages of 3D CNNs over 2D CNNs within the framework of residual learning. Furthermore, we show that factorizing the 3D convolutional filters into separate spatial and temporal components yields significantly gains in accuracy. Our empirical study leads to the design of a new spatiotemporal convolutional block 'R(2+1)D' which produces CNNs that achieve results comparable or superior to the state-of-the-art on Sports-1M, Kinetics, UCF101, and HMDB51.

UR - http://www.scopus.com/inward/record.url?scp=85055123128&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055123128&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2018.00675

DO - 10.1109/CVPR.2018.00675

M3 - Conference contribution

T3 - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

SP - 6450

EP - 6459

BT - Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018

PB - IEEE Computer Society

ER -