Research map / Vision-language and multimodal / Cross-modal learning
Mind the Gap: Aligning Vision Foundation Models to Image Feature Matching
Yuhan Liu, Jingwen Fu, Yang Wu, Kangyi Wu, Pengna Li, Jiayi Wu, Sanping Zhou, Jingmin Xin
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 |
|---|---|---|---|
| use Stable Diffusion features instead of Swin (B) (ablation) | IMD · IMIM | matching accuracy (%): 74.0 -> 88.7 | Improved |
| use Stable Diffusion features instead of DINOv2 (B) (ablation) | IMD · IMIM | matching accuracy (%): 75.5 -> 88.7 | Improved |
| use Stable Diffusion features instead of DINOv2 (B) (ablation) | IMD · MegaDepth | AUC@5: 57.8 -> 61.2 | Improved |
| use Stable Diffusion features instead of Swin (B) (ablation) | IMD · MegaDepth | AUC@5: 57.5 -> 61.2 | Improved |
| add CIPM prompt instead of an empty string (ablation) | IMD · IMIM | matching accuracy (%): 84.0 -> 88.7 | Improved |
| add CIPM prompt instead of an empty string (ablation) | IMD · MegaDepth | AUC@5: 58.4 -> 61.2 | Improved |
| use cross-image prompt instead of individual image prompt (ablation) | IMD · IMIM | matching accuracy (%): 85.2 -> 88.7 | Improved |
| use cross-image prompt instead of individual image prompt (ablation) | IMD · MegaDepth | AUC@5: 59.6 -> 61.2 | Improved |
| add cross attention in CIPM instead of concatenation with shared prompt (ablation) | IMD · IMIM | matching accuracy (%): 87.4 -> 88.7 | Improved |
| add cross attention in CIPM instead of concatenation with shared prompt (ablation) | IMD · MegaDepth | AUC@5: 60.7 -> 61.2 | Improved |
| extract features at t = 0 instead of t = 100 (ablation) | IMD · IMIM | matching accuracy (%): 88.1 -> 88.7 | Improved |
| extract features at t = 0 instead of t = 100 (ablation) | IMD · MegaDepth | AUC@5: 60.9 -> 61.2 | Improved |
Closest papers by meaning
- SD4Match: Learning to Prompt Stable Diffusion Model for Semantic Matching (CVPR 2024 · similarity 0.87)
- OmniGlue: Generalizable Feature Matching with Foundation Model Guidance (CVPR 2024 · similarity 0.86)
- MATCHA: Towards Matching Anything (CVPR 2025 · similarity 0.86)
- $\text{I}^2\text{AM}$: Interpreting Image-to-Image Latent Diffusion Models via Bi-Attribution Maps (ICLR 2025 · similarity 0.85)
- I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models (ICML 2025 · similarity 0.85)
- InstructCV: Instruction-Tuned Text-to-Image Diffusion Models as Vision Generalists (ICLR 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.
