Research map / Detection and segmentation / Open-set and few-shot detection
Reimagining Anomalies: What If Anomalies Were Normal?
Philipp Liznerski, Saurabh Varshneya, Ece Calikus, Puyu Wang, Alexander Bartscher, Sebastian Josef Vollmer, Sophie Fellenz, Marius Kloft
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| train supervised classifier on only blue anomalies instead of unsupervised OE | BCE detector · Colored-MNIST | AuROC: 98% -> 75% | Hurt |
| explain with counterfactuals instead of FCDD heatmaps | anomaly explanation · Colored-MNIST | qualitative coverage | Improved |
| generate CEs for BCE OE detector on CIFAR-10 | counterfactual generator · CIFAR-10 (single normal class) | AuROC: 47.5 | Improved |
| generate CEs for HSC/BCE on MNIST | counterfactual generator · MNIST (single normal class) | AuROC: 72.0 / 80.8 | Hurt |
| explain DSVDD instead of BCE/HSC OE | DSVDD detector · Colored-MNIST | CE quality | Hurt |
| use diffusion-based (DiffEdit) generator for CEs | counterfactual generator · ImageNet-Neighbors | AuROC: 69.1 / 67.9 | Mixed |
Closest papers by meaning
- Unseen Visual Anomaly Generation (CVPR 2025 · similarity 0.88)
- Learning Memory-Guided Normality for Anomaly Detection (CVPR 2020 · similarity 0.88)
- One-to-Normal: Anomaly Personalization for Few-shot Anomaly Detection (NeurIPS 2024 · similarity 0.88)
- Beyond Normal References: Discriminative Few-Shot Anomaly Detection (NeurIPS 2026 · similarity 0.86)
- AnomalyGPT: Detecting Industrial Anomalies Using Large Vision-Language Models (AAAI 2024 · similarity 0.86)
- MetaUAS: Universal Anomaly Segmentation with One-Prompt Meta-Learning (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
