Research map / Label-efficient and robust learning / OOD and anomaly detection
DisCoPatch: Taming Adversarially-driven Batch Statistics for Improved Out-of-Distribution Detection
Francisco Caetano, Christiaan Viviers, Luis A. Zavala-Mondragón, Peter H.N. De With, Fons van der Sommen
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 |
|---|---|---|---|
| use test-batch statistics in BatchNorm instead of learned running statistics (BS=64) (ablation) | DisCoPatch · Near-OOD | AUROC: 38.4 -> 99.3 | Improved |
| use test-batch statistics in BatchNorm instead of learned running statistics (BS=64) (ablation) | DisCoPatch · Far-OOD | AUROC: 34.5 -> 96.1 | Improved |
| increase batch size of test batch from 1 to 256 (ablation) | DisCoPatch · Near-OOD | AUROC: 64.9 -> 100.0 | Improved |
| increase batch size of test batch from 1 to 256 (ablation) | DisCoPatch · Far-OOD | AUROC: 70.9 -> 98.1 | Improved |
| use BatchNorm instead of GroupNorm (ablation) | DisCoPatch-64 · Covariate Shift | AUROC: 78.5 -> 97.2 | Improved |
| use BatchNorm instead of InstanceNorm (ablation) | DisCoPatch-64 · Covariate Shift | AUROC: 86.2 -> 97.2 | Improved |
| use BatchNorm instead of InstanceNorm (ablation) | DisCoPatch-64 · Near-OOD | AUROC: 92.8 -> 95.1 | Improved |
| use BatchNorm instead of GroupNorm (ablation) | DisCoPatch-64 · Far-OOD | AUROC: 81.0 -> 96.6 | Improved |
| increase patches per image from 4 to 64 (ablation) | DisCoPatch · Near-OOD | AUROC: 90.3 -> 95.1 | Improved |
| increase patches per image from 4 to 64 (ablation) | DisCoPatch · Far-OOD | AUROC: 92.6 -> 96.6 | Improved |
| increase patches per image from 64 to 512 (ablation) | DisCoPatch · Near-OOD | AUROC: 95.1 -> 95.3 | No clear effect |
| increase patches per image from 4 to 64 (ablation) | DisCoPatch · Covariate Shift | AUROC: 93.0 -> 97.2 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- RankFeat — RankFeat: Rank-1 Feature Removal for Out-of-distribution Detection (NeurIPS 2022) · better on 6 of 6 shared results
- MOODv2 — MOOD: Multi-Level Out-of-Distribution Detection (CVPR 2021) · better on 4 of 6 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
- Splitting & Integrating: Out-of-Distribution Detection via Adversarial Gradient Attribution (ICML 2025 · similarity 0.90)
- ImageNet-OOD: Deciphering Modern Out-of-Distribution Detection Algorithms (ICLR 2024 · similarity 0.89)
- Leveraging Perturbation Robustness to Enhance Out-of-Distribution Detection (CVPR 2025 · similarity 0.89)
- CADRef: Robust Out-of-Distribution Detection via Class-Aware Decoupled Relative Feature Leveraging (CVPR 2025 · similarity 0.88)
- Mysteries of the Deep: Role of Intermediate Representations in Out of Distribution Detection (NeurIPS 2025 · similarity 0.88)
- Image Background Serves as Good Proxy for Out-of-distribution Data (ICLR 2024 · similarity 0.88)
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.
