Research map / Label-efficient and robust learning / OOD and anomaly detection
Registration Based Few-Shot Anomaly Detection
Chaoqin Huang, Haoyan Guan, Aofan Jiang, Ya Zhang, Michael Spratling, Yan-Feng Wang
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 support set augmentation (ablation) | RegAD · MVTec AD image-level, k=2 | AUC: 79.1 -> 85.7 | Improved |
| add support set augmentation (ablation) | RegAD · MPDD image-level, k=2/4/8 | AUC: 1.2%, 0.5%, 0.6% | Improved |
| add feature registration aggregated training across categories (ablation) | RegAD · MVTec AD image-level, k=2/4/8 | AUC: 3.3%, 2.9%, 2.6% | Improved |
| add spatial transformer network (STN) (ablation) | RegAD · MVTec AD image-level, k=8 | AUC: 89.3% to 91.2% | Improved |
| add spatial transformer network (STN) (ablation) | RegAD · MPDD image-level, k=8 | AUC: 64.8% to 71.9% | Improved |
| add spatial transformer network (STN) (ablation) | RegAD · MVTec AD pixel-level, k=2 | AUC: 94.7 -> 94.6 | No clear effect |
| use affine STN instead of no STN (ablation) | RegAD · MPDD image-level, k=2 | AUC: 52.8 -> 63.4 | Improved |
| use rotation+scale STN instead of no STN (ablation) | RegAD · MVTec AD image-level, k=2 | AUC: 83.0 -> 85.7 | Improved |
| aggregated multi-category training instead of individual training (ablation) | RegAD · MVTec AD image-level, k=2 | AUC: 81.5 -> 85.7 | Improved |
| add rotation augmentation to support set (individual training) (ablation) | RegAD · MVTec AD image-level, k=2 | AUC: 74.7 -> 79.6 | Improved |
| add graying augmentation to support set (individual training) (ablation) | RegAD · MVTec AD image-level, k=2 | AUC: 74.7 -> 74.9 | No clear effect |
| add translation augmentation to support set (individual training) (ablation) | RegAD · MVTec AD image-level, k=2 | AUC: 74.7 -> 77.4 | Improved |
Closest papers by meaning
- Beyond Normal References: Discriminative Few-Shot Anomaly Detection (NeurIPS 2026 · similarity 0.89)
- One-to-Normal: Anomaly Personalization for Few-shot Anomaly Detection (NeurIPS 2024 · similarity 0.89)
- UniVAD: A Training-free Unified Model for Few-shot Visual Anomaly Detection (CVPR 2025 · similarity 0.88)
- Towards Zero-Shot Anomaly Detection and Reasoning with Multimodal Large Language Models (CVPR 2025 · similarity 0.88)
- Kernel-Aware Graph Prompt Learning for Few-Shot Anomaly Detection (AAAI 2025 · similarity 0.88)
- DZAD: Diffusion-based Zero-shot Anomaly Detection (AAAI 2025 · 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.
