EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A histogram transformer approach using attention-based 3D residual network for human action recognition

Maojin Sun iD, Luyi Sun

PLOS One · 29.12.2025 · 20 (12) · e0333893

DOI: 10.1371/journal.pone.0333893 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 29.12.2025. Verfasst von Maojin Sun und 1 weiteren Autor:innen. Beteiligte Einrichtung: CEICloud Data Storage Technology (Beijing) Co., Ltd., No. 15 Countyard on Kechuang Nine Road in Economic-Technological Development Area, Beijing, Chi…. Hauptthemen laut MEDLINE: Neural Networks, Computer, Attention, Pattern Recognition, Automated. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper proposes a lightweight video action recognition framework that integrates 3D Convolutional Neural Networks (CNNs), the Histogram Transformer Block (HTB), and the Split-Attention Residual Block (SAB), while also introducing Spatiotemporal Tensor Factorization (ST-Factor) technology in an innovative manner. The method first incorporates the HTB module into each computational unit of the AR3D backbone network to leverage local statistical features for improve the granularity of spatiotemporal modeling. Next, the SAB module is introduced into the residual path to utilize dynamic channel re-weighting for optimizing feature selection across dimensions. Finally, the ST-Factor decouples the 4D convolution kernels into independent spatial (H × W) and temporal (T × C) operations, which significantly reducing computational redundancy. Experiments on the UCF101/HMDB51 datasets demonstrate that the proposed method not only maintains real-time inference speed but also outperforms existing state-of-the-art (SOTA) methods in recognition accuracy, providing a new paradigm for video understanding research.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Neural Networks, ComputerAttentionPattern Recognition, AutomatedHumansAlgorithmsVideo Recording

Autor:innen und Institutionen

  • CEICloud Data Storage Technology (Beijing) Co., Ltd., No. 15 Countyard on Kechuang Nine Road in Economic-Technological Development Area, Beijing, China.

ORCID-Profile: Maojin Sun

Zitieren

APA

Sun, M., & Sun, L. (2025). A histogram transformer approach using attention-based 3D residual network for human action recognition. PLOS One, 20(12), e0333893. https://doi.org/10.1371/journal.pone.0333893

Vancouver

Sun M, Sun L. A histogram transformer approach using attention-based 3D residual network for human action recognition. PLOS One. 2025;20(12):e0333893. doi:10.1371/journal.pone.0333893

BibTeX
@article{Sun2025,
  author = {Sun, Maojin and Sun, Luyi},
  title = {A histogram transformer approach using attention-based 3D residual network for human action recognition},
  journal = {PLOS One},
  year = {2025},
  volume = {20},
  number = {12},
  pages = {e0333893},
  doi = {10.1371/journal.pone.0333893}
}
Kostenlos starten →