Research map / 3D vision / Depth and stereo
BlendedMVS: A Large-Scale Dataset for Generalized Multi-View Stereo Networks
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 DTU training data with BlendedMVS | R-MVSNet · Tanks and Temples training set | f_score: 0.475 -> 0.532 | Helped |
| Use blended images instead of rendered images, without augmentation (ablation) | MVSNet · DTU validation set | EPE: 2.99 -> 2.88 | Helped |
| Use blended images instead of input images, without augmentation (ablation) | MVSNet · DTU validation set | EPE: 3.70 -> 2.88 | Helped |
| Add online augmentation to rendered-image training (ablation) | MVSNet · DTU validation set | EPE: 2.99 -> 2.94 | Helped |
| Add online augmentation to input-image training (ablation) | MVSNet · DTU validation set | EPE: 3.70 -> 3.16 | Helped |
| Add online augmentation to blended-image training (ablation) | MVSNet · DTU validation set | EPE: 2.88 -> 2.53 | Helped |
| Use blended images instead of rendered images, without augmentation (ablation) | R-MVSNet · DTU validation set | EPE: 5.54 -> 5.77 | Hurt |
| Use blended images instead of input images, without augmentation (ablation) | R-MVSNet · DTU validation set | EPE: 4.47 -> 5.77 | Hurt |
| Add online augmentation to rendered-image training (ablation) | R-MVSNet · DTU validation set | EPE: 5.54 -> 5.10 | Helped |
| Add online augmentation to input-image training (ablation) | R-MVSNet · DTU validation set | EPE: 4.47 -> 3.86 | Helped |
| Add online augmentation to blended-image training (ablation) | R-MVSNet · DTU validation set | EPE: 5.77 -> 3.95 | Helped |
| Use blended images with photometric augmentation rather than rendered images with augmentation (ablation) | MVSNet · DTU validation set | EPE: 2.94 -> 2.53 | Helped |
Closest papers by meaning
- DeepMVS: Learning Multi-View Stereopsis (CVPR 2018 · similarity 0.91)
- Fast-MVSNet: Sparse-to-Dense Multi-View Stereo With Learned Propagation and Gauss-Newton Refinement (CVPR 2020 · similarity 0.89)
- MVS2D: Efficient Multi-View Stereo via Attention-Driven 2D Convolutions (CVPR 2022 · similarity 0.89)
- MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View Stereo (ICLR 2024 · similarity 0.87)
- MVSMamba: Multi-View Stereo with State Space Model (NeurIPS 2025 · similarity 0.87)
- MVD-Fusion: Single-view 3D via Depth-consistent Multi-view Generation (CVPR 2024 · similarity 0.87)
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.