Research map / Label-efficient and robust learning / OOD and anomaly detection
Anomaly Detection using Score-based Perturbation Resilience
Woosang Shin, Jonghyeon Lee, Taehan Lee, Sangmoon Lee, Jong Pil Yun
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 |
|---|---|---|---|
| use score-based perturbation resilience without pre-trained models instead of the best prior unsupervised method NSA | Ours · MVTec AD localization | pixel-level AUROC (%): 96.3 -> 97.4 | Improved |
| use score-based method instead of RIAD reconstruction method | Ours · MVTec AD localization | pixel-level AUROC (%): 94.2 -> 97.4 | Improved |
| use external knowledge (pre-trained model) like PatchCore | anomaly localization on MVTec AD · MVTec AD localization | pixel-level AUROC (%): 97.4 -> 98.1 | Improved |
| use score-based method instead of NSA on image-level detection | Ours · MVTec AD detection | image-level AUROC (%): 97.2 -> 97.7 | Improved |
| increase number of predictions N from 1 to 3 (ablation) | Ours · MVTec AD | image-level AUROC (%): 97.5 -> 97.7 | Improved |
| increase number of predictions N from 3 to 5 (ablation) | Ours · MVTec AD | image-level AUROC (%): 97.7 -> 97.7 | No clear effect |
| tune the variance scale k per category rather than fixed k = 0.001 (ablation) | Ours · MVTec AD detection | image-level AUROC (%): 97.7 -> 98.9 | Improved |
| use too small or too large perturbation variance scale (ablation) | Ours · perturbed data distribution | qualitative | Hurt |
Closest papers by meaning
- Correcting Deviations from Normality: A Reformulated Diffusion Model for Multi-Class Unsupervised Anomaly Detection (CVPR 2025 · similarity 0.89)
- RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detection (CVPR 2024 · similarity 0.88)
- Generating and Reweighting Dense Contrastive Patterns for Unsupervised Anomaly Detection (AAAI 2024 · similarity 0.88)
- A Diffusion-Based Framework for Multi-Class Anomaly Detection (AAAI 2024 · similarity 0.88)
- ResAD: A Simple Framework for Class Generalizable Anomaly Detection (NeurIPS 2024 · similarity 0.87)
- AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Model (AAAI 2024 · similarity 0.87)
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.
