Research map / 3D vision / Depth and stereo
Regularizing Nighttime Weirdness: Efficient Self-Supervised Monocular Depth Estimation in the Dark
Kun Wang, Zhenyu Zhang, Zhiqiang Yan, Xiang Li, Baobei Xu, Jun Li, Jian Yang
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 PBR (priors-based regularization) to the baseline (ablation) | MonoDepth2 · RobotCar-Night | Sq Rel: 7.451 -> 0.539 | Improved |
| add PBR (priors-based regularization) to the baseline (ablation) | MonoDepth2 · nuScenes-Night | Sq Rel: 42.306 -> 4.127 | Improved |
| add SBM (statistics-based mask) to the baseline (ablation) | MonoDepth2 · RobotCar-Night | Sq Rel: 7.451 -> 5.389 | Improved |
| add SBM (statistics-based mask) to the baseline (ablation) | MonoDepth2 · nuScenes-Night | Sq Rel: 42.306 -> 25.794 | Improved |
| add MCIE (mapping-consistent image enhancement) to the baseline (ablation) | MonoDepth2 · RobotCar-Night | Sq Rel: 7.451 -> 6.735 | Improved |
| add MCIE (mapping-consistent image enhancement) to the baseline (ablation) | MonoDepth2 · nuScenes-Night | Sq Rel: 42.306 -> 40.741 | Improved |
| add MCIE to PBR (ablation) | PBR Only · RobotCar-Night | Sq Rel: 0.539 -> 0.528 | Improved |
| add SBM to PBR + MCIE (ablation) | PBR + MCIE · nuScenes-Night | Sq Rel: 4.005 -> 3.793 | Improved |
| add coordinates image Ip to PBR (ablation) | Full Method · RobotCar-Night | Sq Rel: 0.588 -> 0.520 | Improved |
| add coordinates image Ip to PBR (ablation) | Full Method · nuScenes-Night | Sq Rel: 4.467 -> 3.793 | Improved |
| train with PBR regularization as the only loss, without photometric loss (ablation) | PBR · RobotCar-Night | Sq Rel: 0.5204 -> 3.7608 | Hurt |
| use RobotCar-Day instead of nuScenes-Day depth maps as reference (ablation) | Ours · nuScenes-Night | Sq Rel: 3.7926 -> 4.3069 | No clear effect |
Closest papers by meaning
- See through the Dark: Learning Illumination-affined Representations for Nighttime Occupancy Prediction (NeurIPS 2025 · similarity 0.87)
- NightHaze: Nighttime Image Dehazing via Self-Prior Learning (AAAI 2025 · similarity 0.86)
- Synthetic-to-Real Self-supervised Robust Depth Estimation via Learning with Motion and Structure Priors (CVPR 2025 · similarity 0.86)
- NightCC: Nighttime Color Constancy via Adaptive Channel Masking (CVPR 2024 · similarity 0.85)
- NightRain: Nighttime Video Deraining via Adaptive-Rain-Removal and Adaptive-Correction (AAAI 2024 · similarity 0.85)
- Exploring the Common Appearance-Boundary Adaptation for Nighttime Optical Flow (ICLR 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.
