Research map / Vision-language and multimodal / Cross-modal learning
CrossFeat: Bridging Imaging Modalities in Feature Descriptor Space
Paul Schneider, Nazim Haouchine
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 |
|---|---|---|---|
| cross SIFT descriptors with CrossFeat (plus NN matching) | SIFT+NN · Medical (sparse) | SR@1 (%): 11.2 -> 47.0 | Improved |
| cross SIFT descriptors with CrossFeat | SIFT+NN · Driving (sparse) | SR@1 (%): 9.4 -> 73.5 | Improved |
| cross SIFT descriptors with CrossFeat | SIFT+NN · Satellite (sparse) | SR@1 (%): 0.0 -> 34.0 | Improved |
| add test-time adaptation (TTA) after crossing (ablation) | Cross(SIFT)+NN · Medical (sparse) | SR@1 (%): 47.0 -> 79.2 | Improved |
| add test-time adaptation (TTA) after crossing (ablation) | Cross(SIFT)+NN · Satellite (sparse) | SR@1 (%): 34.0 -> 18.0 | Hurt |
| add geometry-appearance disentanglement (ablation) | CrossFeat · ReMIND ablation | SR@1 (%): 39.6 -> 52.9 | Improved |
| add contrastive InfoNCE loss (ablation) | CrossFeat · ReMIND ablation | SR@1 (%): 15.1 -> 52.9 | Improved |
| add adversarial modality loss (ablation) | CrossFeat · ReMIND ablation | SR@1 (%): 40.6 -> 52.9 | Improved |
| add FiLM-conditioned decoder (ablation) | CrossFeat · ReMIND ablation | SR@1 (%): 40.9 -> 52.9 | Improved |
| add variational appearance head (ablation) | CrossFeat · ReMIND ablation | SR@1 (%): 48.3 -> 52.9 | Improved |
| add geometry alignment loss L_geom (ablation) | CrossFeat · ReMIND ablation | SR@1 (%): 48.9 -> 52.9 | Improved |
| use dual-anchor keypoint sampling instead of random (ablation) | CrossFeat · ReMIND ablation | matches per image: 11.2 -> 91.2 | Improved |
Closest papers by meaning
- CrossOver: 3D Scene Cross-Modal Alignment (CVPR 2025 · similarity 0.87)
- Self-Supervised Spatial Correspondence Across Modalities (CVPR 2025 · similarity 0.86)
- When to Align, When to Predict: A Phase Diagram for Multimodal Learning (NeurIPS 2026 · similarity 0.86)
- X-Drive: Cross-modality Consistent Multi-Sensor Data Synthesis for Driving Scenarios (ICLR 2025 · similarity 0.84)
- Connect, Collapse, Corrupt: Learning Cross-Modal Tasks with Uni-Modal Data (ICLR 2024 · similarity 0.84)
- XFeat: Accelerated Features for Lightweight Image Matching (CVPR 2024 · similarity 0.84)
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.
