Explanation of cam_m and cam_c

[Ahmad-Jarrar]

Hi, I checkout out your code. I see that image is multiplied by cam_m and cam_c. However I saw no mention of these parameters in the paper.

Can you explain the purpose and intuition of these parameters.

Thank you.

[weihan1]

i believe they are mentioned in the paper:

"Thus we naively model these differences by simply optimizing a scale and offset parameter for each colour channel for each camera separately."

cam_m and cam_c are the scale and offset parameters that you optimize for the first timestep to account for different variations in color representations