Research map / 3D vision / Depth and stereo
Single-Shot Monocular RGB-D Imaging Using Uneven Double Refraction
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use uneven double refraction with polarizer-controlled intensity imbalance and joint reconstruction | Monocular double-refraction RGB-D imaging · Synthetic dataset | output color PSNR (dB): 26.71 -> 36.63 | Helped |
| Use uneven double refraction with polarizer-controlled intensity imbalance and joint reconstruction | Monocular double-refraction RGB-D imaging · Synthetic dataset | depth RMSE (mm): 212 -> 116 | Helped |
| Use the proposed rectification and GPU reconstruction pipeline | Monocular double-refraction RGB-D imaging · Synthetic dataset | avg. runtime (ms): 38000 -> 34 | Helped |
| Add rectification (ablation) | Double-refraction reconstruction · Synthetic dataset | avg. runtime (ms): 38000 -> 27000 | Helped |
| Add rectification (ablation) | Double-refraction reconstruction · Synthetic dataset | depth RMSE (mm): 212 -> 193 | Helped |
| Add uneven double refraction (ablation) | Rectified double-refraction reconstruction · Synthetic dataset | output color PSNR (dB): 26.80 -> 33.23 | Helped |
| Add uneven double refraction (ablation) | Rectified double-refraction reconstruction · Synthetic dataset | depth RMSE (mm): 193 -> 445 | Hurt |
| Add joint reconstruction (ablation) | Rectified reconstruction with uneven double refraction · Synthetic dataset | output color PSNR (dB): 33.23 -> 36.63 | Helped |
| Add joint reconstruction (ablation) | Rectified reconstruction with uneven double refraction · Synthetic dataset | depth RMSE (mm): 445 -> 116 | Helped |
| Add joint reconstruction (ablation) | Rectified reconstruction with uneven double refraction · Synthetic dataset | avg. runtime (ms): 27000 -> 34 | Helped |
| Use sparse depth estimates instead of ground-truth dense depth for object detection training | FrustumNet v1 · 3D object detection | mAP: 0.86 -> 0.74 | Hurt |
Closest papers by meaning
- High-Quality Stereo Image Restoration From Double Refraction (CVPR 2021 · similarity 0.88)
- Learned Binocular-Encoding Optics for RGBD Imaging Using Joint Stereo and Focus Cues (CVPR 2025 · similarity 0.86)
- Neural RGB(r)D Sensing: Depth and Uncertainty From a Video Camera (CVPR 2019 · similarity 0.86)
- Passive Snapshot Coded Aperture Dual-Pixel RGB-D Imaging (CVPR 2024 · similarity 0.84)
- RGB-to-Polarization Estimation: A New Task and Benchmark Study (NeurIPS 2025 · similarity 0.83)
- Glossy Object Reconstruction with Cost-effective Polarized Acquisition (CVPR 2025 · similarity 0.83)
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.