Unexpected hist_equal result

[Nakilon]

I have such image:

I process it like this and see that it has been splitted into 14 regions:

img.>(65).labelregions.to_a.flatten.uniq
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]

Since they are dark, to see them I apply the hist_equal, but instead of seeing something like this:

It drops several colors:

img.>(65).labelregions.hist_equal.to_a.flatten.uniq
[9, 10, 11, 12]

Is it correct? Is there something I don't understand?

[jcupitt]

Hi @Nakilon,

I think that's correct -- the normalization step in equalization can join very nearby values together.

I usually scale and falsecolour to visualize region labelleling, for example:

x = Vips::Image.new_from_file("regions.png")
labelled, opts = x.labelregions(segments: TRUE)
vis = (labelled * 256 / opts["segments"]).falsecolour

To make:

nip2 has a button to do this (very handy).

[Nakilon]

Cool. Didn't know about falsecolor.