Anomaly Detection2021en
MVTec 3D Anomaly Detection Dataset
4,147 high-resolution 3D point cloud scans and RGB images across 10 categories. First comprehensive 3D anomaly detection benchmark.
Current State of the Art
CMDR-IAD
Research
97.3
auroc
MVTec 3D-AD — auroc
6 results · 3 SOTA advances · higher is better
All results
SOTA frontier
auroc Progress Over Time
Showing 3 breakthroughs from Mar 2022 to Mar 2026
Key Milestones
Total Improvement
12.5%
Time Span
4y 1m
Breakthroughs
3
Current SOTA
97.3
Top Models Performance Comparison
Top 6 models ranked by auroc
Best Score
97.3
Top Model
CMDR-IAD
Models Compared
6
Score Range
10.8
aurocPrimary
| # | Model | Score | Paper / Code | Date |
|---|---|---|---|---|
| 1 | CMDR-IADOpen Source Research | 97.3 | Mar 2025 | |
| 2 | MTSJMOpen Source Research | 95.7 | Mar 2025 | |
| 3 | CFMOpen Source Research | 95.4 | Mar 2025 | |
| 4 | 3D-ADNASOpen Source Anonymous (ECCV 2024) | 95.1 | Mar 2025 | |
| 5 | M3DMOpen Source Fudan University / Alibaba | 94.5 | Mar 2023 | |
| 6 | BTF (Back to the Feature)Open Source Weizmann Institute | 86.5 | Mar 2022 |
Related Papers3
Cross-Modal Mapping and Dual-Branch Reconstruction for 2D-3D Multimodal Industrial Anomaly Detection
Mar 2025Models: CMDR-IAD, MTSJM, CFM +1 more
Multimodal Industrial Anomaly Detection via Hybrid Fusion
Mar 2023Models: M3DM
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection
Mar 2022Models: BTF (Back to the Feature)