- https://github.com/Benjamin-Loison
- Joined on
2022-10-16 -
-
Unclear if correlation should in absolute be close to 1 or just close to 1.
Is it the most appropriate, while staying in the use-case context, to compute {min,max}Color only on training images. Do not forget all denoisers.
Could compute proportion of pixels for mean outside largest other denoisers band.
But it may be a significant proportion exceeding not much, so should maybe just plot distribution of pixel…
Disk Image Writer?
Related to Benjamin_Loison/gimp/issues/29.
Could locate such maximums to potentially see visually other outliers.
>>> index = -100; np.partition(npArray.flatten(), index)[index]
27.28
>>> index = 100; np.partition(npArray.flatten(), index)[index]
-26.35
Could be interesting to just add a line with raw images.
Should make similar plot for RAISE flat-field to maybe understand why Rafael images are absorbing in #63.
If the images are normalized, despite inner bounds different distributions have same contrast modification, right?