Research map / 3D vision / Depth and stereo
StereoDRNet: Dilated Residual StereoNet
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 StereoDRNet without refinement instead of PSMNet | StereoDRNet · SceneFlow | EPE: 1.09 -> 0.98 | Helped |
| Add the refinement network (ablation) | StereoDRNet · SceneFlow | EPE: 0.98 -> 0.86 | Helped |
| Use Vortex instead of Pyramid pooling (ablation) | StereoDRNet · SceneFlow | EPE: 1.17 -> 1.13 | Helped |
| Add the second cost-filtering disparity output (ablation) | StereoDRNet with Vortex pooling · SceneFlow | EPE: 1.13 -> 0.99 | Helped |
| Add the third cost-filtering disparity output (ablation) | StereoDRNet with Vortex pooling · SceneFlow | EPE: 0.99 -> 0.98 | Helped |
| Use photometric error alone in refinement (ablation) | StereoDRNet refinement · SceneFlow | EPE: 0.98 -> 1.03 | Hurt |
| Add geometric error input to refinement (ablation) | StereoDRNet refinement · SceneFlow | EPE: 0.98 -> 0.95 | Helped |
| Add both photometric and geometric error inputs (ablation) | StereoDRNet refinement · SceneFlow | EPE: 0.98 -> 0.93 | Helped |
| Co-train the occlusion map with residual disparity (ablation) | StereoDRNet refinement · SceneFlow | EPE: 0.93 -> 0.86 | Helped |
| Replace Vortex pooling with Pyramid pooling in the full model (ablation) | StereoDRNet · SceneFlow | EPE: 0.86 -> 0.96 | Hurt |
| Increase 3D dilation rates from 1 to 1, 2 (ablation) | StereoDRNet cost filtering · SceneFlow | EPE: 1.13 -> 1.03 | Helped |
| Add dilation rate 4 (ablation) | StereoDRNet cost filtering · SceneFlow | EPE: 1.03 -> 0.98 | Helped |
Closest papers by meaning
- Cost Volume Pyramid Based Depth Inference for Multi-View Stereo (CVPR 2020 · similarity 0.91)
- Efficient Edge-Preserving Multi-View Stereo Network for Depth Estimation (AAAI 2023 · similarity 0.89)
- Learning Non-Volumetric Depth Fusion Using Successive Reprojections (CVPR 2019 · similarity 0.89)
- DeepMVS: Learning Multi-View Stereopsis (CVPR 2018 · similarity 0.88)
- Recurrent MVSNet for High-Resolution Multi-View Stereo Depth Inference (CVPR 2019 · similarity 0.88)
- Non-Parametric Depth Distribution Modelling Based Depth Inference for Multi-View Stereo (CVPR 2022 · similarity 0.88)
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.