Closed andycasey closed 7 years ago
See discussion from @seabroke in #30 and comments from Gal M. The assumption of zero correlations between neighbouring pixels is obviously violated. It's unlikely to affect our results, but we should comment on it as a(nother) violated assumption.
And it is yet another reason that our parameter uncertainties are under-estimated. That should be mentioned.
See discussion from @seabroke in #30 and comments from Gal M. The assumption of zero correlations between neighbouring pixels is obviously violated. It's unlikely to affect our results, but we should comment on it as a(nother) violated assumption.