Research map / 3D vision / Depth and stereo
Stereo Magnification With Multi-Layer Images
Taras Khakhulin, Denis Korzhenkov, Pavel Solovev, Gleb Sterkin, Andrei-Timotei Ardelean, Victor Lempitsky
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 |
|---|---|---|---|
| Use scene-adapted, jointly trained layers rather than regularly spaced planes | StereoLayers-4 vs StereoMag-32 · SWORD | PSNR: 24.45 -> 25.95 | Improved |
| Use scene-adapted, jointly trained layers rather than regularly spaced planes | StereoLayers-4 vs StereoMag-32 · RealEstate10K | PSNR: 31.40 -> 32.61 | Improved |
| Use scene-adapted, jointly trained layers rather than regularly spaced planes | StereoLayers-4 vs StereoMag-32 · LLFF | PSNR: 20.67 -> 22.19 | Improved |
| Use RSBg texturing instead of RBg texturing (ablation) | StereoLayers-4 · SWORD | PSNR: 24.96 -> 25.95 | Improved |
| Use RSBg texturing instead of RAW texturing (ablation) | StereoLayers-4 · SWORD | PSNR: 24.90 -> 25.95 | Improved |
| Use bounds interpolation instead of soft-aggregation depth estimation (ablation) | StereoLayers-4 with RSBg · SWORD | PSNR: 24.66 -> 25.95 | Improved |
| Use bounds interpolation with RSBg instead of soft-aggregation with RAW (ablation) | StereoLayers-4 · SWORD | PSNR: 24.30 -> 25.95 | Improved |
| Use bounds interpolation instead of group compositing depth estimation (ablation) | StereoLayers-4 with RSBg · SWORD | PSNR: 25.24 -> 25.95 | Improved |
| Use bounds interpolation with RSBg instead of group compositing with RAW (ablation) | StereoLayers-4 · SWORD | PSNR: 24.90 -> 25.95 | Improved |
| Increase the number of adaptive layers from four to eight (ablation) | StereoLayers · LLFF | PSNR: 22.19 -> 21.58 | Hurt |
| Reduce the number of adaptive layers from four to two (ablation) | StereoLayers · LLFF | PSNR: 22.19 -> 20.78 | Hurt |
| Post-process and coalesce a 32-plane StereoMag representation into four meshes (ablation) | StereoMag · LLFF | PSNR: 20.67 -> 20.29 | Hurt |
Closest papers by meaning
- Learning To Render Novel Views From Wide-Baseline Stereo Pairs (CVPR 2023 · similarity 0.88)
- Zero-Shot Novel View and Depth Synthesis with Multi-View Geometric Diffusion (CVPR 2025 · similarity 0.86)
- Stereo Radiance Fields (SRF): Learning View Synthesis for Sparse Views of Novel Scenes (CVPR 2021 · similarity 0.86)
- Geometry-guided Online 3D Video Synthesis with Multi-View Temporal Consistency (CVPR 2025 · similarity 0.85)
- FreeSplat: Generalizable 3D Gaussian Splatting Towards Free View Synthesis of Indoor Scenes (NeurIPS 2024 · similarity 0.85)
- Volumetric Surfaces: Representing Fuzzy Geometries with Layered Meshes (CVPR 2025 · 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.
