.png){kind=icon}
.png){kind=icon}
Return
Efficient multi-modal image fusion via flow-based model
.png){kind=icon}
PRE
AI
DOI：10.1016/j.inffus.2026.104627.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We present an efficient flow-based latent framework for image fusion, yielding high-quality results with a single sampling step. • We design a multi-task network that dynamically integrates multi-modality priors by concurrently predicting the flow field and modality-specific integration weights. • We propose the Hiro module, which introduces a matrix-based mutual reinforcement attention mechanism to enable deeply fused, context-aware feature interaction.
Keywords:
Image fusion
Diffusion model
Flow-based generative model
Multi-task network
Journal
IF:
15.5
Papers:
4.1K
Citations:
2.7W
Organization
No organization information available