Research map / 3D vision / Depth and stereo
Cost Volume Pyramid Based Depth Inference for Multi-View Stereo
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 |
|---|---|---|---|
| Replace point-cloud iterative refinement with a coarse-to-fine cost volume pyramid | CVP-MVSNet versus Point-MVSNet · DTU | overall (mm): 0.376 -> 0.351 | Helped |
| Replace point-cloud iterative refinement with a coarse-to-fine cost volume pyramid | CVP-MVSNet versus Point-MVSNet · DTU | runtime: 3.04 -> 1.72 | Helped |
| Use a cost volume pyramid instead of Point-MVSNet's point-cloud approach | CVP-MVSNet versus Point-MVSNet · Tanks and Temples intermediate set | Mean f-score: 48.27 -> 54.03 | Helped |
| Use a cost volume pyramid instead of P-MVSNet | CVP-MVSNet versus P-MVSNet · Tanks and Temples intermediate set | Mean f-score: 55.62 -> 54.03 | Hurt |
| Use 3 pyramid levels instead of 2 (ablation) | CVP-MVSNet · DTU | overall: 0.351 -> 0.366 | Hurt |
| Use 4 pyramid levels instead of 2 (ablation) | CVP-MVSNet · DTU | overall: 0.351 -> 0.375 | Hurt |
| Use 5 pyramid levels instead of 2 (ablation) | CVP-MVSNet · DTU | overall: 0.351 -> 0.376 | Hurt |
| Set evaluation pixel interval to 2 instead of 0.5 (ablation) | CVP-MVSNet · DTU | overall (mm): 0.351 -> 0.356 | Hurt |
| Set evaluation pixel interval to 1 instead of 0.5 (ablation) | CVP-MVSNet · DTU | overall (mm): 0.351 -> 0.351 | No clear effect |
| Set evaluation pixel interval to 0.25 instead of 0.5 (ablation) | CVP-MVSNet · DTU | overall (mm): 0.351 -> 0.397 | Hurt |
Closest papers by meaning
- Non-Parametric Depth Distribution Modelling Based Depth Inference for Multi-View Stereo (CVPR 2022 · similarity 0.95)
- Learning Inverse Depth Regression for Multi-View Stereo with Correlation Cost Volume (AAAI 2020 · similarity 0.93)
- StereoDRNet: Dilated Residual StereoNet (CVPR 2019 · similarity 0.91)
- Cascade Cost Volume for High-Resolution Multi-View Stereo and Stereo Matching (CVPR 2020 · similarity 0.91)
- Efficient Edge-Preserving Multi-View Stereo Network for Depth Estimation (AAAI 2023 · similarity 0.90)
- Recurrent MVSNet for High-Resolution Multi-View Stereo Depth Inference (CVPR 2019 · similarity 0.90)
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.