Re: Nuclei segmentation

Posted by Michael Schmid on Mar 19, 2014; 1:29pm
URL: http://imagej.273.s1.nabble.com/Nuclei-segmentation-tp5006953p5006994.html

Hi Chin,

you can try a 'weaker' watershed to avoid oversegmentation. I have uploaded the Adjustable Watershed to the documentation wiki, there you can try with the 'tolerance' parameter.

http://imagejdocu.tudor.lu/doku.php?id=plugin:segmentation:adjustable_watershed:start

Nevertheless, watershed of the Euclidian Distance Map (EDM) works best on roughly circular particles.

For elongated particles, one should have different metrics than the EDM, e.g. the two largest radii in opposite directions from any point.  If anyone has a plugin that does this instead of the EDM, it would certainly help.

Michael
________________________________________________________________
On Mar 19, 2014, at 09:54, chin wrote:

> Hi Michael,
>
> Thank you for the suggestion. I tried to enhance the contrast then did the
> segmentation. The separation
> of connective objects is a little bit improvment, though, those irregular
> objects are still in the wrong split
> position after performing the watershed.
>
> best,
> chin
>
>
>
> --
> View this message in context: http://imagej.1557.x6.nabble.com/Nuclei-segmentation-tp5006953p5006985.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html