Research map / 3D vision / Depth and stereo
ArgMatch: Adaptive Refinement Gathering for Efficient Dense Matching
Yuxin Deng, Kaining Zhang, Linfeng Tang, Jiaqi Yang, Jiayi Ma
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 upsampler U (ablation) | ConvR baseline · MegaDepth | PCK@0.5px: 76.6 -> 79.0 | Improved |
| add match rectifier R (ablation) | ConvR+U · MegaDepth | AUC@5deg: 58.0 -> 59.4 | Improved |
| replace ConvR with offset estimator O (ablation) | ConvR · MegaDepth | AUC@5deg: 56.3 -> 57.0 | Improved |
| full pipeline O+R+U vs ConvR+R+U (ablation) | ArgMatch · MegaDepth | AUC@5deg: 59.4 -> 61.2 | Improved |
| add scale-adaptive sampling (ablation) | ArgMatch · MegaDepth | AUC@5deg: 60.7 -> 61.2 | Improved |
| add content-aware decoding (ablation) | ArgMatch · MegaDepth | AUC@5deg: 59.9 -> 61.2 | Improved |
| add adaptive gated gathering (ablation) | ArgMatch · MegaDepth | AUC@5deg: 60.1 -> 61.2 | Improved |
| add stop-gradient between stages (ablation) | ConvR · MegaDepth | AUC@5deg: 56.3 -> 57.8 | Improved |
| add stop-gradient between stages (ablation) | ArgMatch · MegaDepth | AUC@5deg: 61.2 -> 60.4 | Hurt |
| ArgMatch vs RoMa efficiency | ArgMatch · MegaDepth-1500 | time (ms): 1557 -> 270 | Improved |
| ArgMatch+ vs RoMa at 0.25m,2deg | ArgMatch+ · InLoc DUC1 | pose accuracy: 54.5 -> 58.6 | Improved |
| ArgMatch vs RoMa on IMC 2022 | ArgMatch · Image Matching Challenge 2022 | mAA@10deg: 85.5 -> 83.2 | Hurt |
Closest papers by meaning
- RoMa v2: Harder Better Faster Denser Feature Matching (ECCV 2026 · similarity 0.88)
- Deep Matching Prior: Test-Time Optimization for Dense Correspondence (ICCV 2021 · similarity 0.87)
- Diffusion Model for Dense Matching (ICLR 2024 · similarity 0.86)
- Unifying Feature and Cost Aggregation with Transformers for Semantic and Visual Correspondence (ICLR 2024 · similarity 0.86)
- Dense Match Summarization for Faster Two-view Estimation (CVPR 2025 · similarity 0.86)
- Efficient LoFTR: Semi-Dense Local Feature Matching with Sparse-Like Speed (CVPR 2024 · similarity 0.85)
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.
