Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

About real Datasets #18

Open
Muccul opened this issue Oct 30, 2020 · 4 comments
Open

About real Datasets #18

Muccul opened this issue Oct 30, 2020 · 4 comments

Comments

@Muccul
Copy link

Muccul commented Oct 30, 2020

@kuihua I don't find 'Real200' , would you provide real Datases of 'real200' ?

@Muccul
Copy link
Author

Muccul commented Oct 31, 2020

i used your pretrained model epoch44, on your test datasets, 100H psnr only is 26.904, and test1200 only is 30.70, Why does it not match the table4 in the paper?
Is Table4 trained on respective data sets, not trained on mixdatas?

@kuijiang94
Copy link
Owner

Thanks for your attention. For the test issue, you can refer to the comments in Replicating Paper results using pretrained weights #7. Our proposed model is trained on the mixed dataset.

@Muccul
Copy link
Author

Muccul commented Oct 31, 2020

@kuihua Thanks! i got correct results.

  1. np.array(img.convert('YCbCr'))[:, :, 0] # using this code, psnr is low, img is PIL image
  2. rgb2ycbcr(img)[:, :, 0] # but using this code, psnr is correct, from skimage.color import rgb2ycbcr

And would you provide real Datases of 'real200' ? i could't find it

@Muccul
Copy link
Author

Muccul commented Nov 3, 2020

@kuihua would you provide real Datases of 'real200' ?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants