HAIAFusion: A Hybrid Attention Illumination-Aware Framework for Infrared and Visible Image Fusion (TIM 2025)

Introduction

This is official implementation of "HAIAFusion: A Hybrid Attention Illumination-Aware Framework for Infrared and Visible Image Fusion" with Pytorch.

Tips

The Trained Model is here.

Recommended Environment

• CUDA 11.1
• conda 4.11.0
• Python 3.7.16
• PyTorch 1.9.0
• timm 0.9.7
• tqdm 4.65.0
• pandas 1.3.5

Framework

Structure of CAM PAM and COA

Citation

If you find this repository useful, please consider citing the following paper:

@article{sun2024TIM,
  author={Sun, Yichen and Dong, Mingli and Yu, Mingxin and Zhu, Lianqing},
  journal={IEEE Transactions on Instrumentation and Measurement}, 
  title={HAIAFusion: A Hybrid Attention Illumination-Aware Framework for Infrared and Visible Image Fusion}, 
  year={2025},
  volume={74},
  number={},
  pages={1-19},
  doi={10.1109/TIM.2025.3527534}
  }  

If you have any questions, feel free to contact me (sunyichen@emails.bjut.edu.cn)

Acknowledgements

Parts of this code repository is based on the following works:

• https://github.com/linklist2/PIAFusion_pytorch