Research map / Image generation / Image restoration
A Task-Guided, Implicitly-Searched and Meta-Initialized Deep Model for Image Fusion
Risheng Liu, Zhu Liu, Jinyuan Liu, Xin Fan, Zhongxuan Luo
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 implicit architecture search instead of DARTS (ablation) | TIM · TNO | MI: 2.414 -> 3.285 | Improved |
| add implicit architecture search instead of DARTS (ablation) | TIM · TNO | VIF: 0.759 -> 0.861 | Improved |
| add pretext meta initialization (ablation) | TIM · TNO | MI: 3.215 -> 3.285 | Improved |
| increase meta-initialization inner updates from 4 to 6 (ablation) | TIM · TNO | MI: 3.285 -> 3.073 | Hurt |
| increase meta-initialization inner updates from 2 to 4 (ablation) | TIM · TNO | VIF: 0.839 -> 0.861 | Improved |
| add task guidance versus direct fusion (ablation) | TIM · IVIF | MI: 2.141 -> 3.285 | Improved |
| add task guidance versus direct fusion (ablation) | TIM · medical image fusion | MI: 1.852 -> 2.359 | Improved |
| add task guidance versus direct fusion (ablation) | TIM · Multi-Spectral detection | mAP: 0.447 -> 0.476 | Improved |
| add task guidance versus direct fusion (ablation) | TIM · MFNet segmentation | mIOU: 0.575 -> 0.685 | Improved |
| raise latency weight lambda from 0 to 2 (ablation) | TIM · TNO | MI and time: 3.569 -> 3.416; 0.1532 -> 0.0012 | Mixed |
| use 5-DB instead of 3-DB single operator (ablation) | single-operator network · TNO | MI: 3.465 -> 3.415 | Hurt |
| use 5-RB instead of 3-DC single operator (ablation) | single-operator network · TNO | MI: 3.130 -> 3.398 | Improved |
Closest papers by meaning
- MMAIF: Multi-task and Multi-degradation All-in-One for Image Fusion with Language Guidance (ICCV 2025 · similarity 0.89)
- Task-driven Image Fusion with Learnable Fusion Loss (CVPR 2025 · similarity 0.89)
- One Model for ALL: Low-Level Task Interaction Is a Key to Task-Agnostic Image Fusion (CVPR 2025 · similarity 0.88)
- Image Fusion via Vision-Language Model (ICML 2024 · similarity 0.87)
- Every SAM Drop Counts: Embracing Semantic Priors for Multi-Modality Image Fusion and Beyond (CVPR 2025 · similarity 0.87)
- Task-Customized Mixture of Adapters for General Image Fusion (CVPR 2024 · similarity 0.87)
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.
