Pose-centric motion synthesis through adaptive instance normalization

Olly Hixon-Fisher; Jarek Francik; Dimitrios Makris

Pose-centric motion synthesis through adaptive instance normalization

Research output: Contribution to conference › Paper › peer-review

9 Downloads (Pure)

Abstract

In pose-centric motion synthesis, existing methods often depend heavily on architecture-specific mechanisms to comprehend temporal dependencies. This paper addresses this challenge by introducing the use of adaptive instance normalization layers to capture temporal coherence within pose-centric motion synthesis. We demonstrate the effectiveness of our contribution through state-of-the-art performance in terms of Fréchet Inception Distance (FID) and comparable diversity scores. Evaluations conducted on the CMU MoCap and the HumanAct12 datasets showcase our method‘s ability to generate plausible and high-quality motion sequences, underscoring its potential for diverse applications in motion synthesis.

Original language	English
Publication status	Published - Feb 2025
Event	20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Porto, Portugal Duration: 26 Feb 2025 → 28 Feb 2025

Conference

Conference	20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
Period	26/02/25 → 28/02/25

Bibliographical note

Note: Published in: Proceedings of the 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2025) - Volume 2: VISAPP, pages 39-47. ISSN 2184-4321 ISBN 9789897587283.

Organising Body: Institute for Systems and Technologies of Information, Control and Communication

Keywords

Computer science and informatics

Access to Document

Dimitrios-M-56779-VoRFinal published version, 471 KBLicence: CC BY-NC-ND

https://www.scitepress.org/PublicationsDetail.aspx?ID=mUFH5/ybU+o=&t=1

Pose-centric motion synthesis through adaptive instance normalization
Hixon-Fisher, O., Francik, J. & Makris, D., Feb 2025, Proceedings of the 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2025). SCITEPRESS, Vol. 2: VISAPP. p. 39-47
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Open Access
File

Cite this

@conference{f2a31c83bd7240639c1e8a912f48977d,

title = "Pose-centric motion synthesis through adaptive instance normalization",

abstract = "In pose-centric motion synthesis, existing methods often depend heavily on architecture-specific mechanisms to comprehend temporal dependencies. This paper addresses this challenge by introducing the use of adaptive instance normalization layers to capture temporal coherence within pose-centric motion synthesis. We demonstrate the effectiveness of our contribution through state-of-the-art performance in terms of Fr{\'e}chet Inception Distance (FID) and comparable diversity scores. Evaluations conducted on the CMU MoCap and the HumanAct12 datasets showcase our method{\textquoteleft}s ability to generate plausible and high-quality motion sequences, underscoring its potential for diverse applications in motion synthesis.",

keywords = "Computer science and informatics",

author = "Olly Hixon-Fisher and Jarek Francik and Dimitrios Makris",

note = "Note: Published in: Proceedings of the 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2025) - Volume 2: VISAPP, pages 39-47. ISSN 2184-4321 ISBN 9789897587283. Organising Body: Institute for Systems and Technologies of Information, Control and Communication; 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications ; Conference date: 26-02-2025 Through 28-02-2025",

year = "2025",

month = feb,

language = "English",

}

TY - CONF

T1 - Pose-centric motion synthesis through adaptive instance normalization

AU - Hixon-Fisher, Olly

AU - Francik, Jarek

AU - Makris, Dimitrios

N1 - Note: Published in: Proceedings of the 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2025) - Volume 2: VISAPP, pages 39-47. ISSN 2184-4321 ISBN 9789897587283. Organising Body: Institute for Systems and Technologies of Information, Control and Communication

PY - 2025/2

Y1 - 2025/2

N2 - In pose-centric motion synthesis, existing methods often depend heavily on architecture-specific mechanisms to comprehend temporal dependencies. This paper addresses this challenge by introducing the use of adaptive instance normalization layers to capture temporal coherence within pose-centric motion synthesis. We demonstrate the effectiveness of our contribution through state-of-the-art performance in terms of Fréchet Inception Distance (FID) and comparable diversity scores. Evaluations conducted on the CMU MoCap and the HumanAct12 datasets showcase our method‘s ability to generate plausible and high-quality motion sequences, underscoring its potential for diverse applications in motion synthesis.

AB - In pose-centric motion synthesis, existing methods often depend heavily on architecture-specific mechanisms to comprehend temporal dependencies. This paper addresses this challenge by introducing the use of adaptive instance normalization layers to capture temporal coherence within pose-centric motion synthesis. We demonstrate the effectiveness of our contribution through state-of-the-art performance in terms of Fréchet Inception Distance (FID) and comparable diversity scores. Evaluations conducted on the CMU MoCap and the HumanAct12 datasets showcase our method‘s ability to generate plausible and high-quality motion sequences, underscoring its potential for diverse applications in motion synthesis.

KW - Computer science and informatics

M3 - Paper

T2 - 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

Y2 - 26 February 2025 through 28 February 2025

ER -

Pose-centric motion synthesis through adaptive instance normalization

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Fingerprint

Research output

Pose-centric motion synthesis through adaptive instance normalization

Cite this