The author introduces ReFusion, a unified image fusion framework based on meta-learning, to address challenges in deep learning-based image fusion algorithms by introducing a parameterized loss function dynamically adjusted for different fusion tasks.
Introducing FusionINN, an invertible image fusion framework for efficient generation and decomposition of fused images.
ReFusion introduces a unified image fusion framework based on meta-learning, providing a learnable fusion loss function that adapts to various fusion tasks.