MaskDiff-AD uses reconstruction difficulty of masked coordinates in a diffusion model trained only on nominal data to detect anomalies, with a non-parametric variant and theoretical error guarantees, achieving the best average rank on 18 datasets.
A survey on diffusion models for anomaly detection
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
A single model trained on one normal sample per dataset from nine heterogeneous medical sources achieves state-of-the-art anomaly detection in one-shot universal, full-shot universal, one-shot specialized, and full-shot specialized settings.
citing papers explorer
-
Masked Diffusion Modeling for Anomaly Detection
MaskDiff-AD uses reconstruction difficulty of masked coordinates in a diffusion model trained only on nominal data to detect anomalies, with a non-parametric variant and theoretical error guarantees, achieving the best average rank on 18 datasets.
-
Semantic Iterative Reconstruction: One-Shot Universal Anomaly Detection
A single model trained on one normal sample per dataset from nine heterogeneous medical sources achieves state-of-the-art anomaly detection in one-shot universal, full-shot universal, one-shot specialized, and full-shot specialized settings.