Research map / Vision-language and multimodal / Cross-modal learning
Text-Driven Fusion for Infrared and Visible Images: Achieving Image Scene Adaptation on Hyperbolic Space
Huan Kang, Hui Li, Tianyang Xu, Tao Zhou, Xiaojun Wu, Josef Kittler
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 text guidance during training and testing (ablation) | TEDFusion · TNO | SD: 43.68 -> 44.52 | Improved |
| add text features at training time with attribute-based inference rather than text-only training (ablation) | TEDFusion · TNO | SD: 43.91 -> 44.52 | Improved |
| add the attribute and text gating fusion module (ATGFM) (ablation) | TEDFusion · TNO | SD: 44.17 -> 44.52 | Improved |
| do text-image association in hyperbolic Poincare ball instead of Euclidean space (ablation) | TEDFusion · TNO | SD: 44.24 -> 44.52 | Improved |
| do text-image association in hyperbolic Poincare ball instead of Euclidean space (ablation) | TEDFusion · TNO | AG: 3.21 -> 3.88 | Improved |
| add the adaptive hyperbolic loss L_h (ablation) | TEDFusion · TNO | SD: 42.70 -> 44.52 | Improved |
| add the adaptive hyperbolic loss L_h (ablation) | TEDFusion · TNO | SF: 12.41 -> 10.90 | Hurt |
| increase curvature from c=0.5 to c=1.0 (ablation) | TEDFusion · TNO | SD: 43.19 -> 44.52 | Improved |
| increase curvature from c=0.7 to c=1.0 (ablation) | TEDFusion · TNO | SF: 9.58 -> 10.90 | Improved |
| use TEDFusion instead of Text-IF | Text-IF · RoadScene | SD: 46.70 -> 60.85 | Improved |
| use TEDFusion instead of Text-IF | Text-IF · TNO | SF: 7.88 -> 10.90 | Improved |
| use TEDFusion instead of Text-IF | Text-IF · M3FD detection | mAP@0.5: 49.23 -> 57.37 | Improved |
Closest papers by meaning
- A Hybrid Space Model for Misaligned Multi-modality Image Fusion (AAAI 2026 · similarity 0.89)
- Compositional Entailment Learning for Hyperbolic Vision-Language Models (ICLR 2025 · similarity 0.88)
- HyperET: Efficient Training in Hyperbolic Space for Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.87)
- Every SAM Drop Counts: Embracing Semantic Priors for Multi-Modality Image Fusion and Beyond (CVPR 2025 · similarity 0.86)
- Image Fusion via Vision-Language Model (ICML 2024 · similarity 0.86)
- HEART: Hyperspherical Embedding Alignment via Kent-Representation Traversal in Diffusion Models (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.
