Research map / 3D vision / Depth and stereo
SPE-MVS: Spatial Position Encoding Enhanced Multi-View Stereo with Monocular Depth Priors
Shaoqian Wang, Jiadai Sun, Bosen Hou, Qiang Wang, Bin Fan, Bo Li, 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 PSHF hybrid feature extractor (ablation) | SPE-MVS baseline skeleton · DTU | Overall (mm): 0.298 -> 0.283 | Improved |
| add SPEC spatial-position cost volume (ablation) | SPE-MVS baseline skeleton · DTU | Overall (mm): 0.298 -> 0.288 | Improved |
| add MDGE module (ablation) | SPE-MVS baseline skeleton · DTU | Overall (mm): 0.298 -> 0.286 | Improved |
| add all of PSHF, SPEC and MDGE (ablation) | SPE-MVS baseline skeleton · DTU | Overall (mm): 0.298 -> 0.272 | Improved |
| dual-branch fusion extractor instead of input concatenation (ablation) | PSHF · DTU | Overall (mm): 0.277 -> 0.272 | Improved |
| dual-branch fusion extractor instead of dual-encoder (ablation) | PSHF · DTU | Overall (mm): 0.276 -> 0.272 | Improved |
| use MFE and MDE together instead of MFE only (ablation) | MDGE · DTU | Overall (mm): 0.274 -> 0.272 | Improved |
| remove monocular feature prior from MFE (ablation) | MDGE · DTU | Overall (mm): 0.274 -> 0.277 | Hurt |
| remove monocular depth prior from MDE (ablation) | MDGE · DTU | Overall (mm): 0.276 -> 0.278 | Hurt |
| add MDGE module increases cost (ablation) | SPE-MVS · DTU | time per view (s): 0.42 -> 0.58 | Hurt |
| replace MonoMVSNet with SPE-MVS | MonoMVSNet · DTU | Completeness (mm): 0.243 -> 0.220 | Improved |
| replace MonoMVSNet with SPE-MVS | MonoMVSNet · DTU | Accuracy (mm): 0.313 -> 0.324 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- MonoMVSNet — MonoMVSNet: Monocular Priors Guided Multi-View Stereo Network (ICCV 2025) · better on 4 of 5 shared results
- RRT-MVS — RRT-MVS: Recurrent Regularization Transformer for Multi-View Stereo (AAAI 2025) · better on 3 of 3 shared results
- MVSFormer++ — MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View Stereo (ICLR 2024) · better on 2 of 3 shared results
- GoMVS — GoMVS: Geometrically Consistent Cost Aggregation for Multi-View Stereo (CVPR 2024) · better on 1 of 1 shared result
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
- MVSAnywhere: Zero-Shot Multi-View Stereo (CVPR 2025 · similarity 0.89)
- MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View Stereo (ICLR 2024 · similarity 0.88)
- RRT-MVS: Recurrent Regularization Transformer for Multi-View Stereo (AAAI 2025 · similarity 0.87)
- A Global Depth-Range-Free Multi-View Stereo Transformer Network with Pose Embedding (NeurIPS 2024 · similarity 0.87)
- Multi-view Reconstruction via SfM-guided Monocular Depth Estimation (CVPR 2025 · similarity 0.86)
- Dual-Level Precision Edges Guided Multi-View Stereo with Accurate Planarization (AAAI 2025 · similarity 0.86)
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.
