FusionDN

This is the code of the following paper (tensorflow):

@inproceedings{xu2020aaai,
  title={FusionDN: A Unified Densely Connected Network for Image Fusion},
  author={Xu, Han and Ma, Jiayi and Le, Zhuliang and Jiang, Junjun and Guo, Xiaojie},
  booktitle={Proceedings of the Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI)},
  pages={12484--12491},
  year={2020}
}

It is a unified model for multiple image fusion tasks, including:

1. visible and infrared image fusion
   
2. multi-exposure image fusion
   
3. multi-focus image fusion
   

Framework:

Overall procedure:

Intuitive description of data flow and the process of EWC:

Fused results:

To train:

CUDA_VISIBLE_DEVICES=0,1 python main.py (2 GPUs are needed)

To test:

CUDA_VISIBLE_DEVICES=0 python test_main.py

Tips:

Large files should be downloaded separately, including the following files:

For training:

• Training dataset
  The code to create your own training dataset can be found here.
• vgg16 (put it in 'VGGnet/')
  

For testing:

• Trained model
  

If you have any question, please email to me ([email protected]).