Research map / 3D vision / Depth and stereo
Revisiting Gradient-Based Uncertainty for Monocular Depth Estimation
Julia Hornauer, Amir El-Ghoussani, 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 |
|---|---|---|---|
| gradient-based uncertainty vs post-processing baseline | Monodepth2 Reg-model · KITTI mono | Abs Rel AUSE: 0.044 -> 0.029 | Improved |
| gradient-based uncertainty vs inference-only dropout RMSE | Monodepth2 Reg-model · KITTI mono | RMSE AUSE: 0.854 -> 0.533 | Improved |
| add gradient uncertainty on top of Log-model | Monodepth2 Log-model · KITTI mono | Abs Rel AUSE: 0.039 -> 0.026 | Improved |
| add gradient uncertainty on top of Self-model | Monodepth2 Self-model · KITTI mono | Abs Rel AUSE: 0.030 -> 0.024 | Improved |
| gradient uncertainty vs Log, nUCE | Monodepth2 Log-model · KITTI mono | nUCE: 0.317 -> 0.006 | Improved |
| feature-space flip instead of image flip for reference (ablation) | Monodepth2 · KITTI mono | Abs Rel AUSE: 0.029 -> 0.028 | Improved |
| rotation by 20 degrees instead of flip for reference (ablation) | Monodepth2 · KITTI mono | Abs Rel AUSE: 0.029 -> 0.033 | Hurt |
| diffusion augmentation instead of flip (ablation) | MonoViT · KITTI mono | Abs Rel AUSE: 0.028 -> 0.034 | Hurt |
| feature flip instead of image flip (ablation) | MonoViT · KITTI mono | Abs Rel AUSE: 0.028 -> 0.026 | Improved |
| rotation 20 degrees instead of flip (ablation) | Monodepth2 · NYU | Abs Rel AUSE: 0.062 -> 0.077 | Hurt |
| gradient method vs inference-only dropout on MonoViT stereo, Log-model | MonoViT stereo Log-model · KITTI stereo | Abs Rel AUSE: 0.022 -> 0.026 | Hurt |
| extra backward pass cost vs dropout inference time | Ours · KITTI Monodepth2 | Inference time ms: 53.01 -> 18.84 | Improved |
Closest papers by meaning
- Gradient-Based Uncertainty for Monocular Depth Estimation (ECCV 2022 · similarity 0.95)
- Uncertainty Quantification in Depth Estimation via Constrained Ordinal Regression (ECCV 2022 · similarity 0.91)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.89)
- DME: Unveiling the Bias for Better Generalized Monocular Depth Estimation (AAAI 2024 · similarity 0.88)
- Depth Pro: Sharp Monocular Metric Depth in Less Than a Second (ICLR 2025 · similarity 0.88)
- MonoInstance: Enhancing Monocular Priors via Multi-view Instance Alignment for Neural Rendering and Reconstruction (CVPR 2025 · 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.
