Research map / Detection and segmentation / Open-set and few-shot detection
DRAEM - A Discriminatively Trained Reconstruction Embedding for Surface Anomaly Detection
Vitjan Zavrtanik, Matej Kristan, Danijel Skočaj
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 reconstructive sub-network before discriminative sub-network (ablation) | DRAEM · MVTec AD | image AUROC: 93.9 -> 98.0 | Improved |
| add reconstructive sub-network before discriminative sub-network (ablation) | DRAEM · MVTec AD localization | pixel AP: 62.5 -> 68.4 | Improved |
| add discriminative sub-network to reconstruction (replace SSIM comparison) (ablation) | DRAEM · MVTec AD | image AUROC: 83.9 -> 98.0 | Improved |
| use MSGMS similarity instead of SSIM in the autoencoder detector (ablation) | Recon.-AE · MVTec AD | image AUROC: 83.9 -> 90.7 | Improved |
| use DRAEM two-network design instead of supervised discriminative backbone (ablation) | DRAEM · MVTec AD | pixel AP: 60.7 -> 68.4 | Improved |
| use DTD instead of ImageNet texture source (ablation) | DRAEM · MVTec AD | image AUROC: 97.9 -> 98.0 | No clear effect |
| use random-color regions instead of textures as anomalies (ablation) | DRAEM · MVTec AD | pixel AP: 68.4 -> 56.5 | Hurt |
| use rectangular masks instead of Perlin noise (ablation) | DRAEM · MVTec AD | pixel AP: 68.4 -> 65.1 | Hurt |
| add image augmentation and opacity randomization (ablation) | DRAEM · MVTec AD | pixel AP: 64.3 -> 68.4 | Improved |
| add opacity randomization only (no image augmentation) (ablation) | DRAEM · MVTec AD | pixel AP: 64.3 -> 68.4 | Improved |
| add image augmentation only (no opacity randomization) (ablation) | DRAEM · MVTec AD | pixel AP: 64.3 -> 64.5 | Mixed |
| reduce anomaly source images to as few as 10 textures with augmentation (ablation) | DRAEM · MVTec AD | detection and localization | No clear effect |
Later papers that compared with this method
6 later accepted papers report DRAEM as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- RcAE: Recursive Reconstruction Framework for Unsupervised Industrial Anomaly Detection (AAAI 2026 · better on 2 of 2 shared results)
- A Comprehensive Augmentation Framework for Anomaly Detection (AAAI 2024 · better on 6 of 6 shared results)
- TransFusion -- A Transparency-Based Diffusion Model for Anomaly Detection (ECCV 2024 · better on 1 of 1 shared result)
- Attention-Conditioned Augmentations for Self-Supervised Anomaly Detection and Localization (AAAI 2023 · better on 2 of 2 shared results)
- DeSTSeg: Segmentation Guided Denoising Student-Teacher for Anomaly Detection (CVPR 2023 · better on 2 of 2 shared results)
- PAD: A Dataset and Benchmark for Pose-agnostic Anomaly Detection (NeurIPS 2023 · better on 2 of 2 shared results)
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
- AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Model (AAAI 2024 · similarity 0.85)
- A Diffusion-Based Framework for Multi-Class Anomaly Detection (AAAI 2024 · similarity 0.85)
- Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection (CVPR 2025 · similarity 0.85)
- CostFilter-AD: Enhancing Anomaly Detection through Matching Cost Filtering (ICML 2025 · similarity 0.85)
- Correcting Deviations from Normality: A Reformulated Diffusion Model for Multi-Class Unsupervised Anomaly Detection (CVPR 2025 · similarity 0.84)
- A Comprehensive Augmentation Framework for Anomaly Detection (AAAI 2024 · similarity 0.84)
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.
