Maybe a mistake at raw2raw#L92 ？

[DavideHe]

code at https://github.com/howardyclo/clcc-cvpr21/blob/master/util/raw2raw.py#L92

M_AC = (1 - w[:,None,None]) * I + w[:,None,None] * M_AB
M_BC = (1 - w[:,None,None]) * I + w[:,None,None] * M_BA

if Ill_c be defined by M_AC ，then M_BC should be M_BC = (1 - w[:,None,None]) * M_BA + w[:,None,None] *I

prove as below:

M_BC  = M_BA @ M_AC = M_BA @ ( (1 - w[:,None,None]) * I + w[:,None,None] * M_AB )
## ==>
M_BC  = ( (1 - w[:,None,None]) * M_BA  + w[:,None,None] * M_BA @M_AB )
## M_BA @M_AB closed to I theoretically.
M_BC = (1 - w[:,None,None]) * M_BA + w[:,None,None] *I