Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Bootstrapped Representation Learning for Skeleton-Based Action Recognition

Moliner, Olivier LU orcid ; Huang, Sangxia and Astrom, Kalle LU orcid (2022) 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2022 In IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops 2022-June. p.4153-4163
Abstract

In this work, we study self-supervised representation learning for 3D skeleton-based action recognition. We extend Bootstrap Your Own Latent (BYOL) for representation learning on skeleton sequence data and propose a new data augmentation strategy including two asymmetric transformation pipelines. We also introduce a multi-viewpoint sampling method that leverages multiple viewing angles of the same action captured by different cameras. In the semi-supervised setting, we show that the performance can be further improved by knowledge distillation from wider networks, leveraging once more the unlabeled samples. We conduct extensive experiments on the NTU-60, NTU-120 and PKU-MMD datasets to demonstrate the performance of our proposed method.... (More)

In this work, we study self-supervised representation learning for 3D skeleton-based action recognition. We extend Bootstrap Your Own Latent (BYOL) for representation learning on skeleton sequence data and propose a new data augmentation strategy including two asymmetric transformation pipelines. We also introduce a multi-viewpoint sampling method that leverages multiple viewing angles of the same action captured by different cameras. In the semi-supervised setting, we show that the performance can be further improved by knowledge distillation from wider networks, leveraging once more the unlabeled samples. We conduct extensive experiments on the NTU-60, NTU-120 and PKU-MMD datasets to demonstrate the performance of our proposed method. Our method consistently outperforms the current state of the art on linear evaluation, semi-supervised and transfer learning benchmarks.

(Less)
Please use this url to cite or link to this publication:
author
; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
host publication
Proceedings - 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2022
series title
IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops
volume
2022-June
pages
11 pages
publisher
IEEE Computer Society
conference name
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2022
conference location
New Orleans, United States
conference dates
2022-06-19 - 2022-06-20
external identifiers
  • scopus:85137757589
ISSN
2160-7508
2160-7516
ISBN
9781665487399
DOI
10.1109/CVPRW56347.2022.00460
language
English
LU publication?
yes
id
8f121d1e-204c-4910-a1b0-75f138add4d4
date added to LUP
2022-11-30 11:21:47
date last changed
2024-06-25 21:51:29
@inproceedings{8f121d1e-204c-4910-a1b0-75f138add4d4,
  abstract     = {{<p>In this work, we study self-supervised representation learning for 3D skeleton-based action recognition. We extend Bootstrap Your Own Latent (BYOL) for representation learning on skeleton sequence data and propose a new data augmentation strategy including two asymmetric transformation pipelines. We also introduce a multi-viewpoint sampling method that leverages multiple viewing angles of the same action captured by different cameras. In the semi-supervised setting, we show that the performance can be further improved by knowledge distillation from wider networks, leveraging once more the unlabeled samples. We conduct extensive experiments on the NTU-60, NTU-120 and PKU-MMD datasets to demonstrate the performance of our proposed method. Our method consistently outperforms the current state of the art on linear evaluation, semi-supervised and transfer learning benchmarks.</p>}},
  author       = {{Moliner, Olivier and Huang, Sangxia and Astrom, Kalle}},
  booktitle    = {{Proceedings - 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2022}},
  isbn         = {{9781665487399}},
  issn         = {{2160-7508}},
  language     = {{eng}},
  pages        = {{4153--4163}},
  publisher    = {{IEEE Computer Society}},
  series       = {{IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops}},
  title        = {{Bootstrapped Representation Learning for Skeleton-Based Action Recognition}},
  url          = {{http://dx.doi.org/10.1109/CVPRW56347.2022.00460}},
  doi          = {{10.1109/CVPRW56347.2022.00460}},
  volume       = {{2022-June}},
  year         = {{2022}},
}