Research map / 3D vision / Depth and stereo
Learning Stereo from Single Images
Jamie Watson, Oisin Mac Aodha, Daniyar Turmukhambetov, Gabriel J. Brostow, Michael Firman
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 |
|---|---|---|---|
| train PSMNet on MfS-synthesized stereo pairs instead of Sceneflow | PSMNet · KITTI 2015 | >3px error (%): 5.80 -> 4.75 | Improved |
| train PSMNet on MfS-synthesized stereo pairs instead of Sceneflow | PSMNet · Middlebury | >2px error (%): 34.99 -> 26.42 | Improved |
| train GANet on MfS data instead of Sceneflow | GANet · Middlebury | >2px error (%): 32.20 -> 24.41 | Improved |
| train iResnet on MfS data instead of Sceneflow | iResnet · ETH3D | >1px error (%): 17.90 -> 11.65 | Improved |
| add depth sharpening (ablation) | Ours (PSMNet, MfS) · KITTI 2012 | >3px error (%): 4.75 -> 4.43 | Improved |
| add depth sharpening (ablation) | Ours (PSMNet, MfS) · Middlebury | >2px error (%): 28.16 -> 26.42 | Improved |
| add depth sharpening (ablation) | Ours (PSMNet, MfS) · ETH3D | EPE: 0.47 -> 0.53 | Hurt |
| add background texture filling (ablation) | Ours (PSMNet, MfS) · KITTI 2015 | >3px error (%): 5.27 -> 4.75 | Improved |
| add background texture filling (ablation) | Ours (PSMNet, MfS) · ETH3D | EPE: 0.68 -> 0.53 | Improved |
| use MiDaS instead of DiW as monocular depth network | Ours (PSMNet, MfS) · Middlebury | >2px error (%): 32.96 -> 26.42 | Improved |
| use MiDaS instead of Megadepth as monocular depth network | Ours (PSMNet, MfS) · ETH3D | >1px error (%): 10.41 -> 8.17 | Improved |
| increase MfS training images from 35,000 to 500,000 (MiDaS) (ablation) | Ours (PSMNet, MfS) · Middlebury | >2px error (%): 28.86 -> 26.42 | Improved |
Closest papers by meaning
- On the confidence of stereo matching in a deep-learning era: a quantitative evaluation (TPAMI 2021 · similarity 0.92)
- Guided Stereo Matching (CVPR 2019 · similarity 0.92)
- Zoom and Learn: Generalizing Deep Stereo Matching to Novel Domains (CVPR 2018 · similarity 0.92)
- Unsupervised Adaptation for Deep Stereo (ICCV 2017 · similarity 0.91)
- A Survey on Deep Learning Techniques for Stereo-Based Depth Estimation (TPAMI 2020 · similarity 0.91)
- Stereo Risk: A Continuous Modeling Approach to Stereo Matching (ICML 2024 · similarity 0.89)
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.
