Return
DIFuse: Dual-Domain interactive fusion of Cross-Modality infrared and visible images with decomposition deep network
P
G
J
X
D
DOI:10.1016/j.infrared.2025.106342.png)
Abstract
En 中文
• Proposes a deep interactive network for infrared–visible image fusion with domain-specific decomposition. • Introduces a progressive feature perception module to capture local textures and global structures. • Designs a cross-modal complementary module to enhance feature alignment across modalities. • Constructs an information compensation module to fuse shallow-deep features and preserve details.
Journal
I
IF:
3.4
Papers:
5.8K
Citations:
1.2W
