Research map / Vision-language and multimodal / Cross-modal learning
Unsupervised Multi-Modal Image Registration via Geometry Preserving Image-to-Image Translation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| replace cross-modality similarity measures with learned translation and registration | R + T · Banana dataset, salient objects | registration accuracy (mean landmark distance): 15.8 -> 6.27 | Helped |
| replace cross-modality similarity measures with learned translation and registration | R + T · Banana dataset, full scene | registration accuracy (mean landmark distance): 27.41 -> 6.93 | Helped |
| train both translation-registration compositions rather than either composition alone (ablation) | R + T · Banana dataset | registration accuracy | Helped |
| add bilateral filtering to the smoothness loss (ablation) | R + T · Banana dataset, test | registration accuracy (mean landmark distance): 11.01 -> 6.93 | Helped |
| add bilateral filtering to the smoothness loss (ablation) | R + T · Banana dataset, train | registration accuracy (mean landmark distance): 9.89 -> 6.12 | Helped |
| use both reconstruction and adversarial losses to update both networks (ablation) | R + T · Banana dataset | registration accuracy (mean landmark distance): 22.03 -> 11.01 | Helped |
| use reconstruction loss for R and GAN loss for T (ablation) | R + T · Banana dataset | registration accuracy (mean landmark distance): 11.01 -> 28.15 | Hurt |
| use GAN loss for R and reconstruction loss for T (ablation) | R + T · Banana dataset | registration accuracy (mean landmark distance): 11.01 -> 29.02 | Hurt |
| reduce the translation network's capacity (ablation) | R + T · Banana dataset | registration accuracy | Helped |
| train only the register-first composition (T after R) (ablation) | R + T · Banana dataset | registration accuracy | Hurt |
Closest papers by meaning
- CARL: A Framework for Equivariant Image Registration (CVPR 2025 · similarity 0.87)
- Self-Supervised Spatial Correspondence Across Modalities (CVPR 2025 · similarity 0.86)
- CrossOver: 3D Scene Cross-Modal Alignment (CVPR 2025 · similarity 0.86)
- FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators (ICLR 2024 · similarity 0.85)
- Modality-Agnostic Structural Image Representation Learning for Deformable Multi-Modality Medical Image Registration (CVPR 2024 · similarity 0.85)
- When to Align, When to Predict: A Phase Diagram for Multimodal Learning (NeurIPS 2026 · 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.