Research map / Label-efficient and robust learning / OOD and anomaly detection
Inter-Realization Channels: Unsupervised Anomaly Detection Beyond One-Class Classification
Declan McIntosh, Alexandra Branzan Albu
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 |
|---|---|---|---|
| add span filter to channels (ablation) | InReaCh · MVTec AD image AUROC (10 corrupt) | AUROC: 0.90998 -> 0.91007 | No clear effect |
| add sigma filter to channels (ablation) | InReaCh · MVTec AD pixel AUROC (10 corrupt) | AUROC: 0.95396 -> 0.96431 | Improved |
| add sigma filter to channels (ablation) | InReaCh · MVTec AD (10 corrupt) | recall: 0.84490 -> 0.29873 | Hurt |
| add positional embedding (ablation) | InReaCh · MVTec AD pixel AUROC (10 corrupt) | AUROC: 0.96431 -> 0.96763 | Improved |
| add positional embedding (ablation) | InReaCh · MVTec AD image AUROC (10 corrupt) | AUROC: 0.91674 -> 0.92327 | Improved |
| add local aggregator (ablation) | InReaCh · MVTec AD pixel AUROC (10 corrupt) | AUROC: 0.95395 -> 0.95396 | No clear effect |
| add 10 anomalous images to training set | InReaCh · MVTec AD image AUROC | AUROC: 0.903 -> 0.923 | Improved |
| quadruple anomalous training images from 10 to 40 | InReaCh · MVTec AD pixel AUROC | AUROC: 0.968 -> 0.960 | Hurt |
| add 40 corrupt training images to PatchCore | PatchCore · MVTec AD image AUROC | AUROC: 0.991 -> 0.728 | Hurt |
| increase association depth up to 10 (ablation) | InReaCh · MVTec AD (10 corrupt) | AUROC | Improved |
| increase minimum channel span (ablation) | InReaCh · MVTec AD (10 corrupt) | AUROC | Hurt |
| raise positional embedding weight too high (ablation) | InReaCh · MVTec AD (10 corrupt) | image AUROC | Hurt |
Closest papers by meaning
- CNC: Cross-modal Normality Constraint for Unsupervised Multi-class Anomaly Detection (AAAI 2025 · similarity 0.87)
- ResAD: A Simple Framework for Class Generalizable Anomaly Detection (NeurIPS 2024 · similarity 0.87)
- Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection (CVPR 2025 · similarity 0.87)
- Correcting Deviations from Normality: A Reformulated Diffusion Model for Multi-Class Unsupervised Anomaly Detection (CVPR 2025 · similarity 0.86)
- MambaAD: Exploring State Space Models for Multi-class Unsupervised Anomaly Detection (NeurIPS 2024 · similarity 0.86)
- A Diffusion-Based Framework for Multi-Class Anomaly Detection (AAAI 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.
