EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Spatial–Temporal Feature Fusion with Multi-Head Attention for Supervised Anomaly Localization

Saba Ramzan, Sajid Ullah Khan, Farrukh Arslan, Arfan Jaffar, Muhammad Naeem, Muhammad Javaid Iqbal

Sensors · 12.11.2026 · 26 (22) · 7153

DOI: 10.3390/s26227153 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 12.11.2026. Verfasst von Saba Ramzan und 5 weiteren Autor:innen. Die Arbeit nennt 51 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Muhammad Naeem, Department of Computer Science and Engineering, Mälardalen University, Västerås, Sweden.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Localizing anomalous activities in surveillance video requires spatial appearance to be interpreted within its temporal context. This study presents a supervised dual-branch architecture that combines YOLOv8n features from a key frame with R3D-18 features from a 16-frame RGB clip. Channel projection and spatial alignment enable convolutional fusion of the two representations. Residual multi-head self-attention then models relationships among spatial positions in the fused feature map. Dense prediction heads estimate activity classes, objectness scores, and bounding boxes on the key frame. On the six categories of UCF-Crime2Local, the reference run without nonmaximum suppression (NMS) achieves 54.21% precision, 57.85% recall, and 53.44% mAP50. Across three training seeds, mean no-NMS mAP50 is 53.44±0.46%. The finalized pipeline with class-aware NMS at IoU 0.5 achieves 54.10±0.50% mAP50 across the same seeds. Values following ± denote sample standard deviations. The complete model exceeds the fusion-without-attention variant by 4.44 mAP percentage points under the common experimental protocol. Class-wise analysis, controlled component comparisons, repeated runs, and video-level uncertainty analysis characterize the model within the specified key-frame localization task. The accompanying implementation documents temporal sampling, annotation validation, optimization, checkpoint selection, and prediction export. Continuous-video evaluation derives event intervals from category-specific frame scores, achieving 57.78% event F1 at temporal IoU 0.3 and a median onset delay of 0.80 s among matched events. These intervals are obtained through temporal postprocessing rather than a learned boundary-prediction head. Cross-dataset generalization and live-stream deployment remain to be evaluated.

Förderung

  • Muhammad Naeem, Department of Computer Science and Engineering, Mälardalen University, Västerås, Sweden

Zitieren

APA

Ramzan, S., Khan, S. U., Arslan, F., Jaffar, A., Naeem, M., & Iqbal, M. J. (2026). Spatial–Temporal Feature Fusion with Multi-Head Attention for Supervised Anomaly Localization. Sensors, 26(22), 7153. https://doi.org/10.3390/s26227153

Vancouver

Ramzan S, Khan SU, Arslan F, Jaffar A, Naeem M, Iqbal MJ. Spatial–Temporal Feature Fusion with Multi-Head Attention for Supervised Anomaly Localization. Sensors. 2026;26(22):7153. doi:10.3390/s26227153

BibTeX
@article{Ramzan2026,
  author = {Ramzan, Saba and Khan, Sajid Ullah and Arslan, Farrukh and Jaffar, Arfan and Naeem, Muhammad and Iqbal, Muhammad Javaid},
  title = {Spatial–Temporal Feature Fusion with Multi-Head Attention for Supervised Anomaly Localization},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {22},
  pages = {7153},
  doi = {10.3390/s26227153}
}
Kostenlos starten →