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

Abstract

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.

Dual-domain self-supervised feature alignment via spectral–spatial representation learning for deepfake anomaly detection
Yi He
PLOS One11 Sep 2026Peer-reviewedCC BY 4.0Full textBioinformatics & computational biology ProbabilityPLOS
Comparison of anomaly detection and supervised learning for acute myeloid leukemia blast identification: A proof-of-concept study
Mezya Sezen, Pavinap Priyaa, Krishnapriya Vinod +4
arXiv29 Oct 2026PreprintarXiv licenceFull textGeneral biologyarXiv
PatHoST: weakly supervised reconstruction of pathogen spatial distributions from host transcriptomes
F. Di Marco, F. Nicola, G. Saldarini +4
bioRxiv26 Oct 2026PreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
FateDriver: Weakly supervised dual-view graph learning for transition-specific prioritization of candidate regulators in cellular reprogramming
J. Sung, K. Son, H.-J. Cha +2
bioRxiv27 Oct 2026PreprintCC BY-NC-NDFull textBioinformatics & computational biologybioRxiv
Text guidance is powerful but prompt-sensitive for weakly-supervised leaf symptom segmentation
Romane Dubois, Lydia Bousset, Stéphane Jumel +3
Smart Agricultural Technology10 Jul 2026Peer-reviewedCC BY 4.0Full textPlant biologyHAL

Cite

BibTeX RIS

BibTeX

@article{zhou2026learn,
  title = {{Learn the interactions: Weakly supervised video anomaly detection with human-object interactions}},
  author = {Maowen Zhou and Erma Rahayu Mohd Faizal Abdullah and Aznul Qalid Md Sabri and Susanto Rahardja},
  journal = {PLOS One},
  year = {2026},
  month = sep,
  doi = {10.1371/journal.pone.0358538},
  url = {https://doi.org/10.1371/journal.pone.0358538},
  note = {CC BY 4.0},
}

RIS

TY  - JOUR
TI  - Learn the interactions: Weakly supervised video anomaly detection with human-object interactions
AU  - Maowen Zhou
AU  - Erma Rahayu Mohd Faizal Abdullah
AU  - Aznul Qalid Md Sabri
AU  - Susanto Rahardja
PY  - 2026
DA  - 2026/09/18
JO  - PLOS One
DO  - 10.1371/journal.pone.0358538
UR  - https://doi.org/10.1371/journal.pone.0358538
AB  - 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.
SN  - 1932-6203
ER  -

Abstract

Related

Cite