EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Learn the interactions: Weakly supervised video anomaly detection with human-object interactions

Maowen Zhou iD, Erma Rahayu Mohd Faizal Abdullah iD, Aznul Qalid Md Sabri iD, Susanto Rahardja

PLOS One · 18.09.2026 · 21 (9) · e0358538

DOI: 10.1371/journal.pone.0358538 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 18.09.2026. Verfasst von Maowen Zhou und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Artificial Intelligence, Universiti Malaya, Kuala Lumpur, Wilayah Persekutuan Kuala Lumpur, Malaysia. Hauptthemen laut MEDLINE: Video Recording, Supervised Machine Learning. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Weakly supervised video anomaly detection remains a challenging problem, primarily due to the scarcity of abnormal training samples and the lack of diverse feature representations, which hamper the learning of discriminative models. To address these issues, we introduce a novel weakly supervised cross-domain framework that enriches feature representations by explicitly integrating semantic cues from Human-Object Interactions (HOI) and action dynamics. Specifically, we model domain-specific features using two independent encoders and fuse them via a bi-directional cross-attention mechanism to obtain a unified embedding. To further alleviate class imbalance and enhance feature discrimination, our network is jointly optimized with classification loss, a cross-domain contrastive loss, and a clip-level focal loss. The proposed method is evaluated with various visual backbones, demonstrating its adaptability and robustness. Extensive experiments show that our approach achieves competitive performance against state-of-the-art methods under weak supervision in terms of AUC. Specifically, our proposed method achieves 98.33% AUC and 87.96% AUC on two benchmark datasets, i.e., ShanghaiTech and UCF-Crime, respectively.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Video RecordingSupervised Machine LearningHumansAlgorithmsDetection Algorithms

Autor:innen und Institutionen

  • Department of Artificial Intelligence, Universiti Malaya, Kuala Lumpur, Wilayah Persekutuan Kuala Lumpur, Malaysia.
  • Engineering Cluster, Singapore Institute of Technology, Singapore, Singapore.

ORCID-Profile: Maowen Zhou, Erma Rahayu Mohd Faizal Abdullah, Aznul Qalid Md Sabri

Zitieren

APA

Zhou, M., Faizal Abdullah, E. R. M., Sabri, A. Q. M., & Rahardja, S. (2026). Learn the interactions: Weakly supervised video anomaly detection with human-object interactions. PLOS One, 21(9), e0358538. https://doi.org/10.1371/journal.pone.0358538

Vancouver

Zhou M, Faizal Abdullah ERM, Sabri AQM, Rahardja S. Learn the interactions: Weakly supervised video anomaly detection with human-object interactions. PLOS One. 2026;21(9):e0358538. doi:10.1371/journal.pone.0358538

BibTeX
@article{Zhou2026,
  author = {Zhou, Maowen and Faizal Abdullah, Erma Rahayu Mohd and Sabri, Aznul Qalid Md and Rahardja, Susanto},
  title = {Learn the interactions: Weakly supervised video anomaly detection with human-object interactions},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {9},
  pages = {e0358538},
  doi = {10.1371/journal.pone.0358538}
}
Kostenlos starten →