1 Star 0 Fork 104

于斌 / PaddleGAN

forked from PaddlePaddle / PaddleGAN 
加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
克隆/下载
prenet.md 2.58 KB
一键复制 编辑 原始数据 按行查看 历史
wangna11BD 提交于 2022-05-12 22:25 . fix prenet doc (#631)

PReNet

1 Introduction

"Progressive Image Deraining Networks: A Better and Simpler Baseline" provides a better and simpler baseline deraining network by considering network architecture, input and output, and loss functions.

2 How to use

2.1 Prepare dataset

The dataset(RainH.zip) used by PReNet can be downloaded from here,uncompress it and get two folders(RainTrainH、Rain100H).

The structure of dataset is as following:

    ├── RainH
        ├── RainTrainH
        |    ├── rain
        |    |    ├── 1.png
        |    |    └── 2.png
        |    |        .
        |    |        .
        |    └── norain
        |        ├── 1.png
        |        └── 2.png
        |            .
        |            .
        └── Rain100H
            ├── rain
            |    ├── 001.png
            |    └── 002.png
            |        .
            |        .
            └── norain
                ├── 001.png
                └── 002.png
                    .
                    .

2.2 Train/Test

train model:

   python -u tools/main.py --config-file configs/prenet.yaml

test model:

   python tools/main.py --config-file configs/prenet.yaml --evaluate-only --load ${PATH_OF_WEIGHT}

3 Results

Evaluated on RGB channels, scale pixels in each border are cropped before evaluation.

The metrics are PSNR / SSIM.

Method Rain100H
PReNet 29.5037 / 0.899

Input:

Output:

4 Model Download

model dataset
PReNet RainH.zip

References

@inproceedings{ren2019progressive,
   title={Progressive Image Deraining Networks: A Better and Simpler Baseline},
   author={Ren, Dongwei and Zuo, Wangmeng and Hu, Qinghua and Zhu, Pengfei and Meng, Deyu},
   booktitle={IEEE Conference on Computer Vision and Pattern Recognition},
   year={2019},
 }
Python
1
https://gitee.com/ybinu/PaddleGAN.git
git@gitee.com:ybinu/PaddleGAN.git
ybinu
PaddleGAN
PaddleGAN
develop

搜索帮助