Hi Dani,
there were a few bugs in the 'Segmented Particles' output of Find Maxima.
According to the documentation, the "Segmented Particles" output should do
a Watershed segmentation. This means segmentation lines, but no other
areas with background color. There should be one exception: If you select
'above lower threshold' everything below that threshold becomes a
background and is excluded from segmentation.
One of the bugs I mentioned was taking the lowest pixel value as
background even without any threshold. This was unintended and it has
caused undesired results with many images (usually, there are only a few
pixels having the lowest value). For example, the number of segmented
particles was not necessarily equal to the number of maxima (caused by
maxima below the threshold, but surrounded by pixels of the lowest value).
I guess that your images have a fixed value of all background pixels. In
such a case (which is not the typical one, at least in my experience) the
bug has caused all background pixels in the image to appear as a
background in the segmented output.
You can restore the old behavior by setting the lower threshold to the
background value plus one (don't care about the upper threshold, it can
be, e.g. the maximum pixel value or the same as the lower threshold) and
checking "Above Lower Threshold". In case your particles are minima of the
pixel value, not maxima, you have to invert the image before "Find
Maxima".
Hope this helps?
Michael
__________________________________________________________________
On Sun, November 15, 2009 13:42, danibodor wrote:
> Hi all,
>
> Recently i have been having a problem with the Find Maxima (Output =
> Segmented Particles) option. It is no longer giving me segmented
> particles,
> but a voronoi diagram seeded with the maximum points. I know that i can
> use
> the voronoi in combination with the original 8bit picture to get the
> segmented picture, but it is a bit stupid to go through all those extra
> steps.
>
> Does anyone know what has happened? i recently updated my ImageJ version
> from 1.43i to 1.41k; and i saw in the documentation that there was an
> update
> in this function:
>>> Michael Schmid fixed a bug in Process>Binary>Find Maxima, with
>>> "Segmented" output, that caused it to >> create extra invalid features.
>
> Did this somehow change the output type altogether?
>
>
> And secondly, i don't know if there is a way to de-update my imageJ
> version
> back to 1.43i. is there?
>
>
>
> Thanks a lot in advance!!
>
> Dani
> --
> View this message in context:
>
http://n2.nabble.com/Find-Maxima-Segmented-Particles-tp4007423p4007423.html> Sent from the ImageJ mailing list archive at Nabble.com.
>
Locked
