search

nothinglo / Deep-Photo-Enhancer

TensorFlow implementation of the CVPR 2018 spotlight paper, Deep Photo Enhancer: Unpaired Learning for Image Enhancement from Photographs with GANs
MIT License
794 stars 110 forks source link

Finally, I got the correct data format and color space. Consistent with the paper data (PSNR) and the imgs from this web #91

Closed TriLoo closed 5 years ago

TriLoo commented 5 years ago

I export all images to .tif format, with sRGB color space.

all steps is same as #38 data export steps. The only difference is the Color Space was set to sRGB.

Use the 27.5 SL model, I got PSNR 23.80023, this is consistent with the paper data good. And my inference outputs looks very close to the images from this repo.