Thanks for release the code

[shendu-ops]

Hello the authors,

Your work is very cool and I am trying to follow your work. While reproducing the code, I noticed that there are two JSON files (gt_illumination.json and sampled_illumination.json) in the ./data_scripts folder. How were these two files generated? Additionally, if I want to build my own dataset, do I need to obtain the ground truth illumination for each RAW image?

[heouchen]

suggest use cursor or other AI IDE to anaysis these questions

[DY112]

Hi, thank you for your interest in my work.
And, sorry for the late reply. I don't know why I didn't get the alert for this issue. I might have missed it :<

gt_illumination includes the ground-truth illumination chromaticity for each dataset.
sampled_illumination includes the augmented illumination chromaticity, following the polynomial augmentation method (from Mahmoud's paper, you can find the exact method in Fig.10 and Sec.D of the paper's supplementary material).

For the second question, yes, you need to know the ground-truth illumination chroma for each raw image so that you can do white-balance and convert to XYZ colorspace, etc.

If you have additional questions, feel free to ask more!