search

csjunxu / Guided-Image-Denoising-TIP2018

External Prior Guided Internal Prior Learning for Real-World Noisy Image Denoising. IEEE Transactions on Image Processing, 2018.
Other
32 stars 21 forks source link

Questions regarding the guided_EI/guided_NoGT #1

Open TouqeerAhmad opened 6 years ago

TouqeerAhmad commented 6 years ago

Hi Jun,

Thank you for sharing the code and data sets!

Can you please explain on a high level what is the difference between 'Denoising_Guided_EI' and 'Denoising_Guided_NoGT' -- is it just you are providing the sigma i.e. the noise level in the first one? I also see '% PGPD denoising' and '% Guided denoising' in the comments, can you please explain what is PGPD -- by guided I understand external prior guiding the internal prior.

Also, in one of the demos 'NoiseEstimation' is called but in others it is not.

Looking forward to hear from you.

Thanks, Touqeer

csjunxu commented 6 years ago

Hi, Touqeer,

The function _NoGT deal with noisy images with no ground truth and hence we do not calculate the PSNR or SSIM in this function. The major objective of the two functions _EI and *_NoGT is the same, i.e., using external prior to guide the learning of internal prior and perform denoising accordingly.

Hope my answer can help you.

Thank you for your interests.

All my best wishes!

Jun Xu

Department of Computing

The Hong Kong Polytechnic University

2018-03-29 6:41 GMT+08:00 TouqeerAhmad notifications@github.com:

Hi Jun,

Thank you for sharing the code and data sets!

Can you please explain on a high level what is the difference between 'Denoising_Guided_EI' and 'Denoising_Guided_NoGT' -- is it just you are providing the sigma i.e. the noise level in the first one? I also see '% PGPD denoising' and '% Guided denoising' in the comments, can you please explain what is PGPD -- by guided I understand external prior guiding the internal prior.

Also, in one of the demos 'NoiseEstimation' is called but in others it is not.

Looking forward to hear from you.

Thanks, Touqeer

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/csjunxu/GID_TIP2018/issues/1, or mute the thread https://github.com/notifications/unsubscribe-auth/AI2cQbQX9CtXsmIgBEP8P_5Z3N9ln3L2ks5tjBGPgaJpZM4S_emo .

TouqeerAhmad commented 6 years ago

Thanks for replying Jun!

I am sorry, I am still unclear about the difference. I have noticed different 'Image2PGs' functions have been called in 'Denoising_Guided_El' and 'Denoising_Guided_NoGT'. Also, in 'Denoising_Guided_El', the S and D matrices are being used explicitly which correspond to dictionary and regularization as per the comments in the file.

I am confused as to why D/S are being used in 'Denoising_Guided_El' but not in 'Denoising_Guided_NoGT'? Since, these are coming from the trained mat file i.e. the external dictionary from the clean images, why it is not being used in both? Can you please explain a bit more about it?

Thanks, Touqeer

TouqeerAhmad commented 6 years ago

Does the 'Denoising_Guided_NoGT' correspond to the "Internal" as referred in the paper specifically in Table 1. As I have also noticed the time improves when I call 'Denoising_Guided_El'