ReDegNet
ReDegNet copied to clipboard
csxmli2016
From Face to Natural Image: Learning Real Degradation for Blind Image Super-Resolution (ECCV 2022)
Results
1
ReDegNet issues
Sort by
recently updated
recently updated
newest added
作者您好，这是一篇很有用的工作。 似乎退化模型ReDegNet和画质修复模型f2n(F2NESRGAN)是可以独立训练的。请问为什么训练代码里面他们是共同训练的？ 之前也有人提出类似的疑惑，他说去掉之后，训练效果没有那么好： https://github.com/csxmli2016/ReDegNet/issues/5#issue-1819434797
Metadata
104
Stars
8
Forks
Watchers
Owner
Metadata
From Face to Natural Image: Learning Real Degradation for Blind Image Super-Resolution (ECCV 2022)