Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
ModelZoo
LapSRN_tensorflow
Commits
77faab78
Commit
77faab78
authored
Jun 08, 2023
by
yongshk
Browse files
更新
parent
17d17806
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
27 additions
and
70 deletions
+27
-70
LapSRN_README.md
LapSRN_README.md
+0
-31
README.md
README.md
+27
-39
No files found.
LapSRN_README.md
deleted
100644 → 0
View file @
17d17806
# Tensorflow implementation of the paper "Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution" (CVPR 2017)
This is a Tensorflow implementation using TensorLayer.
Original paper and implementation using MatConNet can be found on their
[
project webpage
](
http://vllab1.ucmerced.edu/~wlai24/LapSRN/
)
.
### Environment
The implementation is tested using python 3.6 and cuda 8.0.
### Download repository:
$ git clone https://github.com/zjuela/LapSRN-tensorflow.git
### Train model
Specify dataset path in config.py file and run:
$ python main.py
The pre-trained model is trained using
[
NTIRE 2017
](
http://www.vision.ee.ethz.ch/ntire17/
)
challenge dataset.
### Test
Run with your test image:
$ python main.py -m test -f TESTIMAGE
Results can be find in folder ./samples/
README.md
View file @
77faab78
#
简介
#
LapSRN
## 模型介绍
LapSRN是一种用于超分辨率重建的神经网络模型
,全称为Laplacian Pyramid Super-Resolution Network。它可以将低分辨率的图像增强到高分辨率,从而提高图像的质量和清晰度。LapSRN模型基于图像金字塔理论,采用多级金字塔结构,通过渐进性的上采样和细节增强,逐步提高图像分辨率
。
LapSRN是一种用于超分辨率重建的神经网络模型
。它可以将低分辨率的图像增强到高分辨率,从而提高图像的质量和清晰度
。
## 模型结构
LapSRN模型主要有两个部分,即拉普拉斯金字塔预测模型和残差学习模型。
拉普拉斯金字塔预测模型通过建立图像金字塔结构,将低分辨率图像逐步上采样到目标分辨率,并通过像素差异来预测高分辨率图像。残差学习模型则用于学习并纠正拉普拉斯金字塔模型中的误差,从而进一步提高图像质量。
LapSRN模型主要有两个部分,即拉普拉斯金字塔预测模型和残差学习模型。
## 数据集
# 测试流程
数据集下载地址:DIV2K
## 安装工具包
https://data.vision.ee.ethz.ch/cvl/DIV2K/
tensorflow1.15版本
[
[tensorflow-1.15.1_dtk22.04.1-cp37-cp37m-manylinux2014_x86_64
](
https://cancon.hpccube.com:65024/file/4/tensorflow/dtk22.04.1/tensorflow-1.15.1_dtk22.04.1-cp37-cp37m-manylinux2014_x86_64.whl
)
]
## 训练及推理
### 环境配置
##
加载环境变量
加载环境变量
```
export PATH={PYTHON3_install_dir}/bin:$PATH
export PATH={PYTHON3_install_dir}/bin:$PATH
export LD_LIBRARY_PATH={PYTHON3_install_dir}/lib:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH={PYTHON3_install_dir}/lib:$LD_LIBRARY_PATH
```
## 下载数据集
数据集下载地址:DIV2K
https://data.vision.ee.ethz.ch/cvl/DIV2K/
## 修改配置文件
```
python
config
.
valid
.
hr_folder_path
=
'/../LapSRN/data/DIV2K_valid_HR/'
config
.
valid
.
lr_folder_path
=
'/../LapSRN/data/DIV2K_train_LR_bicubic/X4/'
```
python依赖安装:
# 运行指令
tensorflow==1.15.1
### 训练
训练命令:
## 训练模型
config.valid.hr_folder_path = '/../LapSRN/data/DIV2K_valid_HR/'
config.valid.lr_folder_path = '/../LapSRN/data/DIV2K_train_LR_bicubic/X4/'
python main.py
```
$ python main.py
```
## 测试
### 测试
测试命令:
使用您的测试图像运行:
python main.py -m test \
-f TESTIMAGE
```
$ python main.py -m test -f TESTIMAGE
```
结果可以在文件夹 ./samples/ 中找到
# 参考
[
https://github.com/zjuela/LapSRN-tensorflow
](
https://github.com/zjuela/LapSRN-tensorflow
)
## 源码仓库及问题反馈
\ No newline at end of file
*
[
https://github.com/zjuela/LapSRN-tensorflow
](
https://github.com/zjuela/LapSRN-tensorflow
)
## 参考
*
[
https://github.com/zjuela/LapSRN-tensorflow
](
https://github.com/zjuela/LapSRN-tensorflow
)
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment