Research map / Label-efficient and robust learning / OOD and anomaly detection
PNI : Industrial Anomaly Detection using Position and Neighborhood Information
Jaehyeok Bae, Jae-Han Lee, Seyun Kim
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 neighborhood-conditioned MLP distribution to PatchCore-style baseline (ablation) | PNI · MVTec AD | average I-AUROC: 98.92 -> 99.44 | Improved |
| add neighborhood-conditioned MLP distribution (ablation) | PNI · MVTec AD | average P-AUROC: 98.18 -> 98.62 | Improved |
| add position histogram information (ablation) | PNI · MVTec AD (object) | P-AUROC: 98.79 -> 99.04 | Improved |
| add position histogram information (ablation) | PNI · MVTec AD (texture) | I-AUROC: 99.55 -> 99.46 | No clear effect |
| add pixel-wise refinement network (ablation) | PNI · MVTec AD (texture) | P-AUROC: 98.33 -> 98.72 | Improved |
| add pixel-wise refinement network (ablation) | PNI · MVTec AD | average P-AUROC: 98.80 -> 98.98 | Improved |
| use PNI (position, neighborhood, refinement) instead of PatchCore | PatchCore · MVTec AD | I-AUROC: 99.1 -> 99.56 | Improved |
| use PNI instead of PatchCore | PatchCore · MVTec AD | false-positive and false-negative images: 19 FP, 23 FN -> 7 FP, 12 FN | Improved |
| ensemble of three feature extractors in PNI | PNI · MVTec AD | P-AUROC: 98.98 -> 99.06 | Improved |
Closest papers by meaning
- RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detection (CVPR 2024 · similarity 0.86)
- AnomalyNCD: Towards Novel Anomaly Class Discovery in Industrial Scenarios (CVPR 2025 · similarity 0.85)
- ADPretrain: Advancing Industrial Anomaly Detection via Anomaly Representation Pretraining (NeurIPS 2025 · similarity 0.85)
- InCLAD: A Continual Learning Benchmark for Industrial Visual Anomaly Detection (NeurIPS 2026 · similarity 0.84)
- AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Model (AAAI 2024 · similarity 0.84)
- Supervised Anomaly Detection for Complex Industrial Images (CVPR 2024 · similarity 0.84)
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.
