MDSR
MDSR 的 Keras 实现
Install / Use
/learn @foamliu/MDSRREADME
超分辨率网络
MDSR 的 Keras 实现。
原理
请参照论文 Enhanced Deep Residual Networks for Single Image Super-Resolution。
本代码参照了原作者的 Torch 实现:NTIRE2017 和 jmiller656 的 Tensorflow 实现 EDSR-Tensorflow.
MDSR (多尺度模型。 我们提供尺寸x2,x3,x4的模型):

依赖
数据集

按照 说明 下载 ImageNet 数据集。# 超分辨率网络
MDSR 的 Keras 实现。
原理
请参照论文 Enhanced Deep Residual Networks for Single Image Super-Resolution。
本代码参照了原作者的 Torch 实现:NTIRE2017 和 jmiller656 的 Tensorflow 实现 EDSR-Tensorflow.
MDSR (多尺度模型。 我们提供尺寸x2,x3,x4的模型):

依赖
数据集

按照 说明 下载 ImageNet 数据集。
如何使用
训练
$ python train.py
如果想可视化训练效果,请运行:
$ tensorboard --logdir path_to_current_dir/logs

演示
下载预训练的 MDSR模型,放入 "models" 目录。然后执行:
$ python demo.py
|x|输入|输入x4|x2输出|x3输出|x4输出|真实x4|
|---|---|---|---|---|---|---|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|37.19029|36.78687|35.39293|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|35.60590|35.68226|34.58508|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|31.41228|31.59771|30.77240|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|32.29662|32.28041|31.17987|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|40.28474|40.25364|39.15853|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|31.16240|31.33199|30.57106|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|29.95088|30.27624|29.53507|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|33.98172|34.14947|33.39684|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|29.73082|29.90029|29.14516|100|
|图片|
|
|
|
|
|
|
|PSNR|n/a|n/a|35.37303|35.24584|34.36166|100|
评估
在 4268 张验证集图片上测得 PSNR 并求均值:x2=33.57876 dB, x3=33.70763 dB, x4=32.75656 dB。
$ python evaluate.py
Related Skills
proje
Interactive vocabulary learning platform with smart flashcards and spaced repetition for effective language acquisition.
YC-Killer
2.7kA library of enterprise-grade AI agents designed to democratize artificial intelligence and provide free, open-source alternatives to overvalued Y Combinator startups. If you are excited about democratizing AI access & AI agents, please star ⭐️ this repository and use the link in the readme to join our open source AI research team.
best-practices-researcher
The most comprehensive Claude Code skills registry | Web Search: https://skills-registry-web.vercel.app
groundhog
400Groundhog's primary purpose is to teach people how Cursor and all these other coding agents work under the hood. If you understand how these coding assistants work from first principles, then you can drive these tools harder (or perhaps make your own!).
