RGB vs BGR

[yathans]

In function negative_inversion(...) , and in various other places:

r_input, g_input, b_input = cv2.split(image_input)

Per cv2 convention, when you read in an image and split, these are actually b, g, r channels, not r, g, b so they are named incorrectly.

But somehow the code seems to be doing "the right thing", not sure why as the gammas in negative_inversion(...) are not treated symmetrically for each channel, and if you first convert the input image to RGB e.g via cv2.cvtColor(image_input, cv2.COLOR_BGR2RGB), or by renaming the channels after the split, the output has wrong colors.

Great script by the way, I really like it.