Research map / Label-efficient and robust learning / Recognition and long-tail
RGB-Multispectral Matching: Dataset, Learning Methodology, Evaluation
Fabio Tosi, Pierluigi Zama Ramirez, Matteo Poggi, Samuele Salti, Stefano Mattoccia, Luigi Di Stefano
Check 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 proxy labels from the additional RGB camera rather than RGB-MS proxy labels (ablation) | PSM · RGB-MS test set | F-AEPE: 2.30 -> 1.59 | Improved |
| Add continuous-formulation MLPs (ablation) | PSM with 2° RGB supervision · RGB-MS test set | F-AEPE: 1.59 -> 3.03 | Hurt |
| Add the high-resolution RGB feature extractor Φ (ablation) | PSM + MLPs with 2° RGB supervision · RGB-MS test set | F-AEPE: 3.03 -> 1.32 | Improved |
| Collapse RGB and MS inputs to single-channel representations (ablation) | PSM + MLPs + Φ with 2° RGB supervision · RGB-MS test set | F-AEPE: 1.32 -> 3.82 | Hurt |
| Add UnrealStereo4K synthetic training data (ablation) | PSM + MLPs + Φ · RGB-MS test set | F-AEPE: 1.32 -> 1.16 | Improved |
| Replace PSM with GWC (ablation) | MLPs + Φ with synthetic training data · RGB-MS test set | F-AEPE: 1.16 -> 1.37 | Hurt |
| Train with photometric loss instead of proxy labels | Zhi et al. on RGB-MS · RGB-MS test set | F-AEPE: 14.22 -> 3.08 | Hurt |
| Add the high-resolution RGB feature extractor Φ (ablation) | PSM + MLPs with 2° RGB supervision · RGB-MS test set | D-AEPE: 19.12 -> 8.36 | Improved |
Closest papers by meaning
- Color Matching Using Hypernetwork-Based Kolmogorov-Arnold Networks (ICCV 2025 · similarity 0.89)
- Deep Material-Aware Cross-Spectral Stereo Matching (CVPR 2018 · similarity 0.88)
- Leveraging Multispectral Sensors for Color Correction in Mobile Cameras (CVPR 2026 · similarity 0.87)
- Discriminative Multi-Modal Feature Fusion for RGBD Indoor Scene Recognition (CVPR 2016 · similarity 0.86)
- RGB-to-Polarization Estimation: A New Task and Benchmark Study (NeurIPS 2025 · similarity 0.84)
- Self-Supervised Spatial Correspondence Across Modalities (CVPR 2025 · 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.
