Research map / Label-efficient and robust learning / OOD and anomaly detection
Out-of-Distribution Detection for Monocular Depth Estimation
Julia Hornauer, Adrian Holzbock, Vasileios Belagiannis
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 |
|---|---|---|---|
| train encoder and decoder as autoencoder instead of reusing depth features (ablation) | image decoder OOD detection (Monodepth2) · KITTI avg OOD | AUROC: 96.02 -> 52.17 | Hurt |
| train encoder and decoder as autoencoder instead of reusing depth features (ablation) | image decoder OOD detection (MonoViT) · KITTI avg OOD | AUROC: 97.71 -> 35.04 | Hurt |
| train image decoder simultaneously with depth model instead of post hoc (ablation) | image decoder (Monodepth2) · KITTI avg OOD | AUROC: 96.02 -> 98.57 | Improved |
| train image decoder simultaneously with depth model instead of post hoc (ablation) | image decoder (MonoViT) · KITTI avg OOD | AUROC: 97.71 -> 94.20 | Hurt |
| train image decoder together with depth model (ablation) | Monodepth2 · KITTI | Abs Rel: 0.090 -> 0.101 | Hurt |
| train image decoder together with depth model (ablation) | MonoViT · KITTI | Abs Rel: 0.079 -> 0.085 | Hurt |
| reconstruction-error OOD score instead of Post | Monodepth2 · KITTI vs vKITTI | AUROC: 75.98 -> 99.40 | Improved |
| reconstruction-error OOD score instead of Post | Monodepth2 · KITTI vs India | AUROC: 59.61 -> 89.29 | Improved |
| reconstruction-error OOD score instead of Drop | Pixelformer · NYU vs Places | AUROC: 88.13 -> 93.18 | Improved |
| reconstruction-error OOD score instead of Log | Monodepth2 · NYU vs Places | AUROC: 77.32 -> 74.84 | Hurt |
| reconstruction-error OOD score instead of Log | Monodepth2 · NYU vs Places | FPR95: 68.00 -> 65.33 | Improved |
| reconstruction-error OOD score instead of Post | MonoViT · KITTI vs India | AUROC: 46.53 -> 95.95 | Improved |
Closest papers by meaning
- Key Feature Replacement of In-Distribution Samples for Out-of-Distribution Detection (AAAI 2023 · similarity 0.87)
- Rethinking Out-of-Distribution (OOD) Detection: Masked Image Modeling Is All You Need (CVPR 2023 · similarity 0.86)
- Density-driven Regularization for Out-of-distribution Detection (NeurIPS 2022 · similarity 0.86)
- Image Background Serves as Good Proxy for Out-of-distribution Data (ICLR 2024 · similarity 0.85)
- RUNA: Object-Level Out-of-Distribution Detection via Regional Uncertainty Alignment of Multimodal Representations (AAAI 2025 · similarity 0.85)
- Rethinking the Evaluation of Out-of-Distribution Detection: A Sorites Paradox (NeurIPS 2024 · similarity 0.85)
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.
