EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Semantically Grounded Reverse Distillation for Unsupervised Anomaly Detection

Chunlei Li, Xin Yuan, Junpu Wang, Ruimin Yang, Hongling Zhao

Electronics · 07.11.2026 · 15 (21) · 5178

DOI: 10.3390/electronics15215178 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 07.11.2026. Verfasst von Chunlei Li und 4 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Knowledge distillation-based unsupervised anomaly detection has achieved strong performance in industrial inspection. However, existing Reverse Distillation (RD) methods suffer from two fundamental limitations: the bottleneck compresses semantically entangled CNN features into a one-class representation that inadequately characterizes normal structural regularity, and the pairwise distillation objective leaves the co-learned teacher-student geometry semantically underconstrained. To address these issues, we propose Semantically Grounded Reverse Distillation (SGRD), which integrates a frozen vision foundation model (DINOv2 in this work) as a fixed semantic reference into the RD framework. Specifically, a foundation-aware semantic bottleneck conditions the bottleneck encoding pathway with foundation-model patch semantics to produce a structurally coherent one-class code. Furthermore, a foundation-semantic representation anchoring module jointly constrains both teacher and student representations toward the foundation-model semantic subspace, discouraging semantically arbitrary representation geometry. Experiments on MVTec AD achieve 98.5% image-level AUROC and 98.3% pixel-level AUROC, while experiments on VisA achieve 99.8% image-level AUROC and 94.2% pixel-level AUPRO, with the proposed method attaining the strongest overall average performance among the compared methods on both benchmarks.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Li, C., Yuan, X., Wang, J., Yang, R., & Zhao, H. (2026). Semantically Grounded Reverse Distillation for Unsupervised Anomaly Detection. Electronics, 15(21), 5178. https://doi.org/10.3390/electronics15215178

Vancouver

Li C, Yuan X, Wang J, Yang R, Zhao H. Semantically Grounded Reverse Distillation for Unsupervised Anomaly Detection. Electronics. 2026;15(21):5178. doi:10.3390/electronics15215178

BibTeX
@article{Li2026,
  author = {Li, Chunlei and Yuan, Xin and Wang, Junpu and Yang, Ruimin and Zhao, Hongling},
  title = {Semantically Grounded Reverse Distillation for Unsupervised Anomaly Detection},
  journal = {Electronics},
  year = {2026},
  volume = {15},
  number = {21},
  pages = {5178},
  doi = {10.3390/electronics15215178}
}
Kostenlos starten →