Research map / 3D vision / Depth and stereo
Deep Depth From Aberration Map
Masako Kashiwagi, Nao Mishima, Tatsuo Kozakaya, Shinsaku Hiura
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 positional branch (attention on image position) to AMA-Net (ablation) | AMA-Net · Stripe pattern, F1.8 | mean depth error (mm): 25.1 -> 10.0 | Improved |
| increase aperture F-number from F1.8 to F4 to F8 | AMA-Net · MSCOCO test images (focus 1500mm) | mean depth error (mm): 10 -> 20 -> 46.3 | Hurt |
| replace stereo SGM with AMA-Net at F4 for long range | depth measurement · Extended range (focus 6500mm) | mean depth error (mm): 251.1 -> 251.2 | No clear effect |
| use F8 instead of stereo | AMA-Net · Extended range (focus 6500mm) | mean depth error (mm): 251.1 -> 635.7 | Hurt |
| apply f=50mm model to f=24mm and f=105mm lenses without retraining | AMA-Net · Other lenses | qualitative depth map | Mixed |
| train and test with an ideal pillbox PSF (no aberration) (ablation) | AMA-Net · Simulation | depth estimation | Hurt |
| add chromatic aberration (RGB PSF instead of mono PSF) (ablation) | AMA-Net · Simulation, on/off-axis PSF | mean depth error | Improved |
| cross-test models trained on center PSF with corner PSF images (ablation) | AMA-Net · Simulation | depth estimation | Hurt |
Closest papers by meaning
- Deep Optics for Monocular Depth Estimation and 3D Object Detection (ICCV 2019 · similarity 0.88)
- Single Image Depth Estimation Trained via Depth From Defocus Cues (CVPR 2019 · similarity 0.87)
- Depth Pro: Sharp Monocular Metric Depth in Less Than a Second (ICLR 2025 · similarity 0.86)
- Depth Any Camera: Zero-Shot Metric Depth Estimation from Any Camera (CVPR 2025 · similarity 0.85)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.85)
- Zero-shot Depth Completion via Test-time Alignment with Affine-invariant Depth Prior (AAAI 2025 · 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.
