Research map / 3D vision / 3D generation
Learning Dense Feature Matching via Lifting Single 2D Image to 3D Space
Yingping Liang, Yutao Hu, Wenqi Shao, Ying Fu
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 3D-aware encoder (Stage 1) (ablation) | L2M · ZEB GL3 | AUC@5: 50.2 -> 51.5 | Improved |
| add 3D-aware encoder (Stage 1) (ablation) | L2M · ZEB BLE | AUC@5: 41.6 -> 46.0 | Improved |
| add 3D-aware encoder (Stage 1) (ablation) | L2M · ZEB SCE | AUC@5: 34.4 -> 38.5 | Improved |
| add 3D-aware encoder (Stage 1) (ablation) | L2M · ZEB MUL | AUC@5: 61.9 -> 61.7 | Hurt |
| train decoder on large-scale synthetic data (Stage 2) instead of MegaDepth only (ablation) | L2M · ZEB GL3 | AUC@5: 46.0 -> 50.2 | Improved |
| train decoder on large-scale synthetic data (Stage 2) instead of MegaDepth only (ablation) | L2M · ZEB ETO | AUC@5: 77.2 -> 83.6 | Improved |
| L2M vs RoMa (outdoor) on zero-shot benchmark | L2M · ZEB mean | AUC@5: 48.8 -> 51.8 | Improved |
| L2M vs RoMa on cross-modal RGB-IR | L2M · METU-VisTIR | AUC@5: 25.61 -> 30.13 | Improved |
| L2M vs RoMa after fine-tuning on MegaDepth | L2M · MegaDepth-1500 | AUC@5: 62.6 -> 63.1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- LightGlue — LightGlue: Local Feature Matching at Light Speed (ICCV 2023) · better on 2 of 2 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
- RoMa v2: Harder Better Faster Denser Feature Matching (ECCV 2026 · similarity 0.88)
- CoMatcher: Multi-View Collaborative Feature Matching (CVPR 2025 · similarity 0.88)
- RoMa: Robust Dense Feature Matching (CVPR 2024 · similarity 0.87)
- Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting (CVPR 2025 · similarity 0.86)
- OmniGlue: Generalizable Feature Matching with Foundation Model Guidance (CVPR 2024 · similarity 0.86)
- DGC-GNN: Leveraging Geometry and Color Cues for Visual Descriptor-Free 2D-3D Matching (CVPR 2024 · 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.
