Research map / 3D vision / Depth and stereo
EC-MVSNet: Enhanced Cascaded Multi-View Stereo with Cross-Scale Relevance Integration
Shaoqian Wang, Jiadai Sun, Bin Fan, Qiang Wang, Bin Lu, Yuchao Dai
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 |
|---|---|---|---|
| add CFC joint cross-scale cost volume (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.303 -> 0.290 | Improved |
| add CPE probability-guided enhancement (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.303 -> 0.284 | Improved |
| add MFR monocular refinement (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.303 -> 0.288 | Improved |
| add CPE on top of CFC (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.290 -> 0.279 | Improved |
| add MFR on top of CFC+CPE (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.279 -> 0.275 | Improved |
| aggregate adjacent higher-scale features instead of lower (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.275 -> 0.278 | Hurt |
| aggregate both higher and lower scale features in CFC (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.275 -> 0.274 | No clear effect |
| aggregate features from all scales in CFC (ablation) | cascade MVS baseline · DTU test | Overall (mm): 0.275 -> 0.274 | No clear effect |
| use EC-MVSNet instead of MVSFormer++ (ablation) | cascade MVS baseline · DTU test | Completeness (mm): 0.252 -> 0.253 | Hurt |
| use EC-MVSNet instead of RRT-MVS | EC-MVSNet · Tanks and Temples Intermediate | mean F1: 68.16 -> 69.32 | Improved |
| use EC-MVSNet instead of RRT-MVS | EC-MVSNet · Tanks and Temples Advanced | mean F1: 43.29 -> 44.63 | Improved |
| use EC-MVSNet instead of GoMVS | EC-MVSNet · ETH3D Testing | Recall: 86.85 -> 84.53 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- GoMVS — GoMVS: Geometrically Consistent Cost Aggregation for Multi-View Stereo (CVPR 2024) · better on 8 of 11 shared results
- RRT-MVS — RRT-MVS: Recurrent Regularization Transformer for Multi-View Stereo (AAAI 2025) · better on 5 of 5 shared results
- MVSFormer++ — MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View Stereo (ICLR 2024) · better on 4 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CL-MVSNet: Unsupervised Multi-View Stereo with Dual-Level Contrastive Learning (ICCV 2023 · similarity 0.91)
- Deep Stereo Using Adaptive Thin Volume Representation With Uncertainty Awareness (CVPR 2020 · similarity 0.89)
- MonoMVSNet: Monocular Priors Guided Multi-View Stereo Network (ICCV 2025 · similarity 0.89)
- Efficient Edge-Preserving Multi-View Stereo Network for Depth Estimation (AAAI 2023 · similarity 0.89)
- MVSAnywhere: Zero-Shot Multi-View Stereo (CVPR 2025 · similarity 0.88)
- MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View Stereo (ICLR 2024 · 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.
