Research map / Label-efficient and robust learning / OOD and anomaly detection
DSR – A Dual Subspace Re-Projection Network 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 |
|---|---|---|---|
| generate anomalies in quantized feature space instead of image space (ablation) | DSR · MVTec | AP_loc: 67.5 -> 70.2 | Improved |
| generate anomalies in quantized feature space instead of image space (ablation) | DSR · MVTec | AUROC_det: 97.8 -> 98.2 | Improved |
| add similarity-bounded codebook sampling (ablation) | DSR · MVTec | AUROC_det: 95.6 -> 98.2 | Improved |
| add similarity-bounded codebook sampling (ablation) | DSR · MVTec | AP_loc: 62.5 -> 70.2 | Improved |
| use both image and feature reconstruction losses (ablation) | DSR · MVTec | AUROC_det: 96.3 -> 98.2 | Improved |
| use both image and feature reconstruction losses (vs feature only) (ablation) | DSR · MVTec | AP_loc: 61.8 -> 70.2 | Improved |
| add the upsampling module (ablation) | DSR · MVTec | AP_loc: 65.2 -> 70.2 | Improved |
| add the upsampling module (ablation) | DSR · MVTec | AUROC_det: 98.2 -> 98.2 | No clear effect |
| vary similarity bound lambda_s from 0.05 to 0.5 (ablation) | DSR · MVTec | AUROC_det: 98.2 -> 95.5 | Hurt |
| use a few segmented anomalous training images (N=16 instead of 0) | DSR · KSDD2 | AP_det: 87.2 -> 91.4 | Improved |
| use full 246 segmented anomalous images instead of none | DSR · KSDD2 | AP_loc: 61.4 -> 85.5 | Improved |
| DSR vs fully supervised baseline at N=246 | DSR · KSDD2 | AP_det: 95.4 -> 95.2 | No clear effect |
Closest papers by meaning
- DRAEM - A Discriminatively Trained Reconstruction Embedding for Surface Anomaly Detection (ICCV 2021 · similarity 0.88)
- Hyperbolic Anomaly Detection (CVPR 2024 · similarity 0.86)
- A Diffusion-Based Framework for Multi-Class Anomaly Detection (AAAI 2024 · similarity 0.85)
- ResAD: A Simple Framework for Class Generalizable Anomaly Detection (NeurIPS 2024 · similarity 0.85)
- A Unified Latent Schrodinger Bridge Diffusion Model for Unsupervised Anomaly Detection and Localization (CVPR 2025 · similarity 0.84)
- MambaAD: Exploring State Space Models for Multi-class Unsupervised Anomaly Detection (NeurIPS 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.
