Research map / Image generation / Image restoration
High-Quality Stereo Image Restoration From 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 |
|---|---|---|---|
| Restore o-ray and e-ray jointly with a two-headed network rather than derive e-ray by subtracting a one-headed o-ray estimate (ablation) | Stereo restoration network · Middlebury 2014 synthetic test | PSNR: 38.75, 35.08 -> 39.81, 36.33 | Mixed |
| Use attenuation ratio τ=0.3 rather than τ=0.15 or τ=0.45 (ablation) | Stereo restoration network · Synthetic double-refraction images | PSNR (dB), o-ray: 36.02, 39.80, and 35.94 dB for τ=0.15, 0.3, and 0.45 | Helped |
| Restore both rays directly from the double-refraction input instead of using depth/deconvolution-based restoration | Stereo restoration network · Middlebury 2014 synthetic test | PSNR (dB), o-ray: 28.27 -> 39.80; 34.19 -> 39.80 | Helped |
| Restore both rays directly from the double-refraction input instead of using depth/deconvolution-based restoration | Stereo restoration network · Middlebury 2014 synthetic test | PSNR (dB), e-ray: 15.99 -> 34.51; 20.10 -> 34.51 | Helped |
| Estimate dense disparity from the restored stereo pair rather than densify previous methods' sparse depth | Stereo restoration network with PSM-Net · Middlebury 2014 synthetic test | Disparity RMSE (px.), dense: 2.96, 2.07 -> 1.71 | Helped |
| Train with attenuation ratio τ=0.3 and evaluate on other attenuation ratios (ablation) | Stereo restoration network · Synthetic double-refraction images at τ∈{0.15, 0.3, 0.45} | PSNR (dB), o-ray: 36.02, 39.80, and 35.94 dB | Mixed |
Closest papers by meaning
- Single-Shot Monocular RGB-D Imaging Using Uneven Double Refraction (CVPR 2020 · similarity 0.88)
- Learned Binocular-Encoding Optics for RGBD Imaging Using Joint Stereo and Focus Cues (CVPR 2025 · similarity 0.85)
- HELVIPAD: A Real-World Dataset for Omnidirectional Stereo Depth Estimation (CVPR 2025 · similarity 0.83)
- Any-Stereo: Arbitrary Scale Disparity Estimation for Iterative Stereo Matching (AAAI 2024 · similarity 0.83)
- MonSter: Marry Monodepth to Stereo Unleashes Power (CVPR 2025 · similarity 0.83)
- Dual Exposure Stereo for Extended Dynamic Range 3D Imaging (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.