search

zhenngbolun / S-Net

S-Net: A Scalable Convolutional Neural Network for JPEG Compression Artifact Reduction
24 stars 5 forks source link

Evaluation code #5

Closed sachinpuranik99 closed 4 years ago

sachinpuranik99 commented 4 years ago

Can you provide the code that you used to evaluate SSIM/PSNR on the WIN143 data?

zhenngbolun commented 4 years ago

use matlab bicubic downsample all images by factor 2，then use jpeg compressing all images and restore these images

On 10/22/2019 01:31, Sachin Puranik wrote:

Can you provide the code that you used to evaluate the WIN143 data?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

zhenngbolun commented 4 years ago

By the way，you can reference our recent work idcn in my github. the relevant results reported in idcn are much more comprehensive.

On 10/22/2019 01:31, Sachin Puranik wrote:

Can you provide the code that you used to evaluate the WIN143 data?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

sachinpuranik99 commented 4 years ago

Thanks, I will refer the IDCN repo. Another question, I tried to download the WIN143 dataset using the link provided in the README, seems like it requires Baidu account. Is it possible to reupload the WIN143 dataset to Google Drive or any other shared drive which does not require an account to download the data?

zhenngbolun commented 4 years ago

ok，I will provide a onedrive link later. You can try it tomorrow.

On 10/22/2019 06:12, Sachin Puranik wrote:

Thanks, I will refer the IDCN repo. Another question, I tried to download the WIN143 dataset using the link provided in the README, seems like it requires Baidu account. Is it possible to reupload the WIN143 dataset to Google Drive or any other shared drive which does not require an account to download the data?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.