From 79799ca96994ea0121dcbca013f4de165d3a4e7b Mon Sep 17 00:00:00 2001 From: wangna11BD <79366697+wangna11BD@users.noreply.github.com> Date: Thu, 12 May 2022 22:25:42 +0800 Subject: [PATCH] fix prenet doc (#631) --- docs/en_US/tutorials/prenet.md | 40 +++++++++++++++++++------------- docs/zh_CN/tutorials/prenet.md | 42 ++++++++++++++++++++-------------- 2 files changed, 49 insertions(+), 33 deletions(-) diff --git a/docs/en_US/tutorials/prenet.md b/docs/en_US/tutorials/prenet.md index fbe141c..b68e8ca 100644 --- a/docs/en_US/tutorials/prenet.md +++ b/docs/en_US/tutorials/prenet.md @@ -16,30 +16,30 @@ The structure of dataset is as following: ``` - ├── data + ├── RainH ├── RainTrainH - ├── rain - ├── 1.png - └── 2.png - . - . - └── norain - ├── 1.png - └── 2.png - . - . + | ├── rain + | | ├── 1.png + | | └── 2.png + | | . + | | . + | └── norain + | ├── 1.png + | └── 2.png + | . + | . └── Rain100H ├── rain - ├── 001.png - └── 002.png - . - . + | ├── 001.png + | └── 002.png + | . + | . └── norain ├── 001.png └── 002.png . . -``` + ``` ### 2.2 Train/Test @@ -55,6 +55,14 @@ ``` ## 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: diff --git a/docs/zh_CN/tutorials/prenet.md b/docs/zh_CN/tutorials/prenet.md index 2f0e98c..65a2b64 100644 --- a/docs/zh_CN/tutorials/prenet.md +++ b/docs/zh_CN/tutorials/prenet.md @@ -15,24 +15,24 @@ Progressive Image Deraining Networks: A Better and Simpler Baseline提出一种 数据集文件结构如下: ``` - ├── data + ├── RainH ├── RainTrainH - ├── rain - ├── 1.png - └── 2.png - . - . - └── norain - ├── 1.png - └── 2.png - . - . + | ├── rain + | | ├── 1.png + | | └── 2.png + | | . + | | . + | └── norain + | ├── 1.png + | └── 2.png + | . + | . └── Rain100H ├── rain - ├── 001.png - └── 002.png - . - . + | ├── 001.png + | └── 002.png + | . + | . └── norain ├── 001.png └── 002.png @@ -53,8 +53,16 @@ Progressive Image Deraining Networks: A Better and Simpler Baseline提出一种 python tools/main.py --config-file configs/prenet.yaml --evaluate-only --load ${PATH_OF_WEIGHT} ``` -## 3 预测结果 +## 3 实验结果展示 +实验数值结果是在 RGB 通道上进行评估,并在评估之前裁剪每个边界的尺度像素。 + +度量指标为 PSNR / SSIM. +| 模型 | Rain100H | +|---|---| +| PReNet | 29.5037 / 0.899 | + +可视化展示: 输入: