# Densefuse: A Fusion Approach to Infrared and Visible Images - Tensorflow
[Hui Li](https://hli1221.github.io/), Xiao-Jun Wu*
Published in: IEEE Transactions on Image Processing
*H. Li, X. J. Wu, “DenseFuse: A Fusion Approach to Infrared and Visible Images,” IEEE Trans. Image Process., vol. 28, no. 5, pp. 2614–2623, May. 2019.*
- [IEEEXplore](https://ieeexplore.ieee.org/document/8580578)
- [arXiv](https://arxiv.org/abs/1804.08361)
## Note
In 'main.py' file, you will find how to run these codes.
The evaluate methods which used in our paper are shown in 'analysis_MatLab'. And these methods are implemented by MatLab.
## Abstract
In this paper, we present a novel deep learning architecture for infrared and visible images fusion problem.
In contrast to conventional convolutional networks, our encoding network is combined by convolutional neural network layer and dense block which the output of each layer is connected to every other layer. We attempt to use this architecture to get more useful features from source images in encoder process. Then appropriate fusion strategy is utilized to fuse these features. Finally, the fused image is reconstructed by decoder.
Compare with existing fusion methods, the proposed fusion method achieves state-of-the-art performance in objective and subjective assessment.
### The framework of fusion method
![](https://github.com/hli1221/imagefusion_densefuse/blob/master/figures/framework.png)
### Fusion strategy - addition
<img src="https://github.com/hli1221/imagefusion_densefuse/blob/master/figures/fuse_addition.png" width="600">
### Fusion strategy - l1-norm
<img src="https://github.com/hli1221/imagefusion_densefuse/blob/master/figures/fuse_l1norm.png" width="600">
## Training
![](https://github.com/hli1221/imagefusion_densefuse/blob/master/figures/train.png)
We train our network using [MS-COCO 2014](http://images.cocodataset.org/zips/train2014.zip)(T.-Y. Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Dollar, and C. L. Zitnick. Microsoft coco: Common objects in context. In ECCV, 2014. 3-5.) as input images which contains 80000 images and all resize to 256×256 and RGB images are transformed to gray ones. Learning rate is 1×10^(-4). The batch size and epochs are 2 and 4, respectively. Our method is implemented with GTX 1080Ti and 64GB RAM.
## Experimental results
### Infrared and visible images('street')
![](https://github.com/hli1221/imagefusion_densefuse/blob/master/figures/fused_street.png)
### Infrared and visible images(RGB)
Database:
Hwang S, Park J, Kim N, et al. Multispectral pedestrian detection: Benchmark dataset and baseline[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015: 1037-1045.
![](https://github.com/hli1221/imagefusion_densefuse/blob/master/figures/ivrgb_results.png)
### Multi-focus images(RGB)
![](https://github.com/hli1221/imagefusion_densefuse/blob/master/figures/fused_color.png)
If you have any question about this code, feel free to reach me(hui_li_jnu@163.com, lihui@stu.jiangnan.edu.cn)
# Citation
*H. Li, X. J. Wu, “DenseFuse: A Fusion Approach to Infrared and Visible Images,” IEEE Trans. Image Process., vol. 28, no. 5, pp. 2614–2623, May 2019.*
```
@article{li2018densefuse,
title={DenseFuse: A Fusion Approach to Infrared and Visible Images},
author={Li, Hui and Wu, Xiao-Jun},
journal={IEEE Transactions on Image Processing},
volume={28},
number={5},
pages={2614--2623},
month={May},
year={2019},
publisher={IEEE}
}
```
## Pytorch version is available at [here](https://github.com/hli1221/densefuse-pytorch) (FOR REFERENCE ONLY)
没有合适的资源?快使用搜索试试~ 我知道了~
imagefusion_densefuse:DenseFuse(IEEE TIP 2019)-TensorFlow 1.8.0
共40个文件
pyc:10个
py:10个
png:7个
需积分: 50 9 下载量 40 浏览量
2021-05-06
14:39:44
上传
评论 1
收藏 29.64MB ZIP 举报
温馨提示
Densefuse:红外和可见图像的融合方法-Tensorflow 吴小军* 发表于:IEEE图像处理事务 H. Li,XJ Wu,“ DenseFuse:红外和可见图像的融合方法”,IEEE Trans。 图像处理。 28号5月,第2614–2623页,5月。 2019。 笔记 在“ main.py”文件中,您将找到如何运行这些代码。 本文中使用的评估方法显示在“ analysis_MatLab”中。 这些方法是由MatLab实现的。 抽象的 在本文中,我们提出了一种针对红外和可见图像融合问题的新型深度学习架构。 与传统的卷积网络相比,我们的编码网络是由卷积神经网络层和密集块组合而成的,密集块的每一层的输出都与其他每一层相连。 我们尝试使用此体系结构从编码器过程中的源图像中获取更多有用的功能。 然后,采用适当的融合策略融合这些特征。 最后,融合图像由解码器重建。 与现有的融合
资源推荐
资源详情
资源评论
收起资源包目录
imagefusion_densefuse-master.zip (40个子文件)
imagefusion_densefuse-master
generate.py 11KB
images
IV_images.zip 5.03MB
MF_images.zip 2.8MB
fusion_addition.py 112B
fusion_l1norm.py 992B
models
densefuse_gray.zip 3.12MB
utils.py 4KB
README.md 4KB
.idea
imagefusion_densefuse.iml 528B
misc.xml 221B
encodings.xml 135B
workspace.xml 14KB
vcs.xml 180B
inspectionProfiles
profiles_settings.xml 228B
modules.xml 294B
__pycache__
ssim_loss_function.cpython-36.pyc 1KB
generate.cpython-36.pyc 7KB
decoder.cpython-36.pyc 2KB
train_recons_single_loss.cpython-36.pyc 5KB
train_recons.cpython-36.pyc 5KB
fusion_addition.cpython-36.pyc 252B
densefuse_net.cpython-36.pyc 1KB
utils.cpython-36.pyc 3KB
fusion_l1norm.cpython-36.pyc 904B
encoder.cpython-36.pyc 3KB
figures
fused_street.png 1.51MB
framework.png 147KB
fused_color.png 3.79MB
fuse_addition.png 28KB
ivrgb_results.png 1.71MB
train.png 85KB
fuse_l1norm.png 133KB
densefuse_net.py 1KB
validation
validation.zip 11.72MB
main.py 3KB
decoder.py 2KB
train_recons.py 7KB
analysis_MatLab
analysis_MatLab.zip 12KB
ssim_loss_function.py 1KB
encoder.py 4KB
共 40 条
- 1
资源评论
janejane815
- 粉丝: 24
- 资源: 4611
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功