search

happycaoyue / Pseudo-ISP

60 stars 10 forks source link

Incorrect Colour and checkerboard artefact in Noisy images #7

Open MrBled opened 2 years ago

MrBled commented 2 years ago

Hi there. Firstly thank you for the very interesting paper! I am very interested in modeling noise profiles with your program. I have trained the model using the default settings, with DND images denoised using a DNCNN network. However, the dataset generated (using DIV2K) does not seem to have noise similar to the original DND images. There seems to be some incorrect colour shifts as well as some checkerboard artefacts.

Did you observe this in your work and have you any advice for me with regards to settings I could change? Does training for longer help? This seems to be a problem in training PseudoISP as my training data appears to be good. I will try to train again. For now below is an example of the images produced.

Checkerboarding

84_clean84_noisy

Colour artefacts 67_clean67_noisy

Strange colour with Bayer/checkerboard artefacts? 42_clean42_noisy

I did try to train DNCNN again using this but it made performance worse. Any suggestions are welcome

happycaoyue commented 2 years ago

I am sorry to have only just replied to your message. I found this discolouration problem when I first started working on PseudoISP and the main solution was to resize the patch size for training PseudoISP.

Hi there. Firstly thank you for the very interesting paper! I am very interested in modeling noise profiles with your program. I have trained the model using the default settings, with DND images denoised using a DNCNN network. However, the dataset generated (using DIV2K) does not seem to have noise similar to the original DND images. There seems to be some incorrect colour shifts as well as some checkerboard artefacts.

Did you observe this in your work and have you any advice for me with regards to settings I could change? Does training for longer help? This seems to be a problem in training PseudoISP as my training data appears to be good. I will try to train again. For now below is an example of the images produced.

Checkerboarding

84_clean84_noisy

Colour artefacts 67_clean67_noisy

Strange colour with Bayer/checkerboard artefacts? 42_clean42_noisy

I did try to train DNCNN again using this but it made performance worse. Any suggestions are welcome

I am sorry to have only just replied to your message. I found this Incorrect Colour problem when I first started working on PseudoISP and the main solution was to resize the patch size for training PseudoISP. If you have any further questions, feel free to contact me.