Research map / 3D vision / Depth and stereo
Physically Grounded Monocular Depth via Nanophotonic Wavefront Encoding
Bingxuan Li, Jiahao Wu, Yuan Xu, Zezheng Zhu, Yunxiang Zhang, Kenneth Chen, Yanqi Liang, Nanfang Yu, Qi Sun
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 pretrained depth prior (initialize from pretrained ViT) (ablation) | Ours-Small with metalens · FlyingThings3D | MAE: 0.061 -> 0.022 | Improved |
| use metalens design instead of DeepDfD optics (ablation) | Ours-Small · FlyingThings3D | MAE: 0.035 -> 0.022 | Improved |
| use ViT instead of U-Net backbone (no prior) (ablation) | Ours-Small with metalens · FlyingThings3D | MAE: 0.063 -> 0.061 | No clear effect |
| fine-tune pretrained ViT with DeepDfD optics instead of U-Net from scratch (ablation) | DeepDfD baseline · FlyingThings3D | MAE: 0.089 -> 0.035 | Improved |
| add decoder-side fusion module to input adaptation (ablation) | Ours-Base · Physical experiment | AbsRel: 0.055 -> 0.058 | Hurt |
| use input adaptation instead of fusion only (ablation) | Ours-Base · Physical experiment | AbsRel: 0.109 -> 0.055 | Improved |
| add disocclusion-aware forward model (ablation) | Ours-Base · Physical experiment (real data included) | AbsRel: 0.074 -> 0.055 | Improved |
| add polarization-aware augmentation (ablation) | Ours-Base · Physical experiment (real data included) | AbsRel: 0.066 -> 0.055 | Improved |
| add polarization-aware augmentation (ablation) | Ours-Base · Physical experiment (synthetic only) | delta_0.5: 0.240 -> 0.546 | Improved |
| add light imbalance augmentation (ablation) | Ours-Base · Physical experiment (synthetic only) | delta_0.5: 0.451 -> 0.546 | Improved |
| add noise augmentation (ablation) | Ours-Base · Physical experiment (real data included) | delta_0.5: 0.907 -> 0.895 | Hurt |
| fine-tune Depth Anything v2 on our dataset without depth encoding | DepthAny. v2 · NYU Depth V2 (simulated) | AbsRel | No clear effect |
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.
- MoGe v2 — MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision (CVPR 2025) · better on 2 of 2 shared results
- Depth Pro — Depth Pro: Sharp Monocular Metric Depth in Less Than a Second (ICLR 2025) · better on 1 of 1 shared result
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
- SynLF: Zero-Shot Metric Depth from Light Field Cameras via Physics-Grounded Synthesis (ECCV 2026 · similarity 0.88)
- Repurposing Marigold for Zero-Shot Metric Depth Estimation via Defocus Blur Cues (NeurIPS 2025 · similarity 0.87)
- UniDepth: Universal Monocular Metric Depth Estimation (CVPR 2024 · similarity 0.85)
- SharpDepth: Sharpening Metric Depth Predictions Using Diffusion Distillation (CVPR 2025 · similarity 0.85)
- Depth Pro: Sharp Monocular Metric Depth in Less Than a Second (ICLR 2025 · similarity 0.85)
- Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data (CVPR 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.
