Research map / Label-efficient and robust learning / OOD and anomaly detection
MATCH: Flow Matching for Multi-View Anomaly Detection
Mathis Kruse, Melissa Schween, Bodo Rosenhahn
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 Hutchinson divergence term (M=1) (ablation) | MATCH · Real-IAD | I-AUROC: 91.172 -> 91.170 | No clear effect |
| add Hutchinson divergence term (M=1) (ablation) | MATCH · Real-IAD | FPS: 18.77 -> 5.40 | Hurt |
| add Hutchinson divergence term (M=4) (ablation) | MATCH · Real-IAD | FPS: 18.77 -> 1.56 | Hurt |
| add divergence term (M=1) (ablation) | MATCH · Real-IAD | memory (GB): 8.88 -> 10.28 | Hurt |
| replace Euler with Midpoint solver (ablation) | MATCH · Real-IAD | I-AUROC: 91.17 -> 91.08 | No clear effect |
| replace Euler with RK4 solver (ablation) | MATCH · Real-IAD | I-AUROC: 91.17 -> 91.07 | No clear effect |
| replace Euler with adaptive Dopri5 (ablation) | MATCH · Real-IAD | FPS: 13.13 -> 0.48 | Hurt |
| increase Euler step size from 0.05 to 0.2 (ablation) | MATCH · Real-IAD | I-AUROC: 91.09 -> 91.17 | No clear effect |
| increase Euler step size from 0.05 to 0.2 (ablation) | MATCH · Real-IAD | FPS: 3.26 -> 13.13 | Improved |
| use flow matching instead of RealNVP normalizing flow | anomaly detector · Real-IAD | I-AUROC: 90.27 -> 91.17 | Improved |
| use flow matching instead of RealNVP normalizing flow | anomaly detector · Real-IAD | bits per dim: 2.10 -> 1.83 | Improved |
| use MATCH instead of RD4AD with matched parameter count | anomaly detector · Real-IAD | I-AUROC: 86.75 -> 91.17 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- Dinomaly — Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection (CVPR 2025) · better on 0 of 1 shared result
- SimpleNet — SimpleNet: A Simple Network for Image Anomaly Detection and Localization (CVPR 2023) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Odd-One-Out: Anomaly Detection by Comparing with Neighbors (CVPR 2025 · similarity 0.87)
- Dual-Interrelated Diffusion Model for Few-Shot Anomaly Image Generation (CVPR 2025 · similarity 0.86)
- CostFilter-AD: Enhancing Anomaly Detection through Matching Cost Filtering (ICML 2025 · similarity 0.86)
- AnomalyGPT: Detecting Industrial Anomalies Using Large Vision-Language Models (AAAI 2024 · similarity 0.86)
- AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Model (AAAI 2024 · similarity 0.86)
- Unveiling Multi-View Anomaly Detection: Intra-view Decoupling and Inter-view Fusion (AAAI 2025 · similarity 0.86)
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.
