Research map / Label-efficient and robust learning / OOD and anomaly detection
MultiADS: Defect-aware Supervision for Multi-type Anomaly Detection and Segmentation in Zero-Shot Learning
Ylli Sadikaj, Hongkuan Zhou, Lavdim Halilaj, Stefan Schmid, Steffen Staab, Claudia Plant
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 KBA defect-aware prompts (ablation) | MultiADS · MTAS MVTec-AD -> VisA | pixel AUROC: 87.0 -> 93.6 | Improved |
| add KBA defect-aware prompts (ablation) | MultiADS · MTAS MVTec-AD -> MAD-sim | pixel AP: 26.5 -> 31.5 | Improved |
| use 4 feature stages instead of 3 (ablation) | MultiADS · ZSAD MVTec-AD -> VisA | image AP: 82.3 -> 86.9 | Improved |
| add KBA prompts to ZSAD (ablation) | MultiADS · ZSAD MVTec-AD -> VisA | image AP: 86.1 -> 86.9 | Improved |
| add global anomaly score a_x (ablation) | MultiADS · ZSAD MVTec-AD -> VisA | image AP: 85.8 -> 86.9 | Improved |
| add KBA prompts to ZSAD on MPDD (ablation) | MultiADS · ZSAD MVTec-AD -> MPDD | image AP: 77.6 -> 78.4 AP; AUROC 77.9 -> 78.3 | Mixed |
| filter product-irrelevant defect types at inference (MultiADS-F) (ablation) | MultiADS · MPDD / MAD-real / Real-IAD ZSAD | image AUROC: MPDD 78.3 -> 79.7; Real-IAD 78.7 -> 78.2 | Mixed |
| train on seen-defect-type data vs unseen defect types | MultiADS · MTAS VisA, trained on MVTec-AD | pixel F1-score / AP: Hole AP 25.19 vs Extra AP 0.15 | Mixed |
| train on MVTec-AD or Real-IAD instead of MPDD to target VisA | MultiADS · MTAS VisA | pixel AP: 23.3 -> 24.8 | Improved |
Closest papers by meaning
- MuSc: Zero-Shot Industrial Anomaly Classification and Segmentation with Mutual Scoring of the Unlabeled Images (ICLR 2024 · similarity 0.89)
- MVREC: A General Few-shot Defect Classification Model Using Multi-View Region-Context (AAAI 2025 · similarity 0.88)
- Supervised Anomaly Detection for Complex Industrial Images (CVPR 2024 · similarity 0.88)
- 3CAD: A Large-Scale Real-World 3C Product Dataset for Unsupervised Anomaly Detection (AAAI 2025 · similarity 0.88)
- Towards Open-Vocabulary Industrial Defect Understanding with a Large-Scale Multimodal Dataset (CVPR 2026 · similarity 0.87)
- AnomalyNCD: Towards Novel Anomaly Class Discovery in Industrial Scenarios (CVPR 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.
