[Insight-users] 3D separation of touching grains

[Richard Beare]

Hi,

There are a couple of factors that influence how well a distance
transform/watershed approach will work - namely the variability in
size of the objects you are interested in, how circular they are and
the degree of overlap between them. If you have a densely sampled 3D
image then I'd imagine that the objects would be touching rather than
overlapping. Overlapping problems tend to arise more often in
brightfield microscopy situations.

The other factor to consider is whether there is information in the
raw image data that might help, for example are there cell nuclei that
don't overlap that would could be used as markers to segment the cell
body.

My rules of thumb are 1) If the overlap is large then a distance
transform based approach won't split the object - in this case you'd
need to think carefully about the measurement you want to make and
whether it makes sense when there is a lot of overlap. 2) If the
objects are not close to circular the a distance transform approach is
likely to over segment.

If the size of the objects don't vary much then you can do various
sorts of filtering to help reduce the chances of over segmentation.

Finally, do you really want to segment? If you only want to count or
estimate size distributions then there are other approaches you might
look at.

On Thu, Apr 17, 2008 at 5:03 PM, Gaëtan Lehmann
<gaetan.lehmann at jouy.inra.fr> wrote:
>
>  Le 17 avr. 08 à 08:46, BOURGEOIS Loic a écrit :
>
>
> > Hi,
> >
> > Thank you for your reply, I had already try it and indeed, it is a very
> efficient method, but I have issues about over-segmentation so I just wanted
> to know if an other method (based on 3D curvature) could be used.
> >
>
>  I don't know how to do that with curvature based methods (I'm interested
> though, so if you find how to do that, please report your success on the
> mailing list).
>
>  However, you can avoid the over segmentation by calling the SetLevel()
> method on the watershed filter. Please look at the Figure 4, page 7 in the
> same article to see the effect of the Level parameter.
>
>  Regards,
>
>  Gaëtan
>
>
>
>
>
> >
> >
> > Regards,
> >
> > Loic
> >
> > -----Message d'origine-----
> > De : Gaëtan Lehmann [mailto:gaetan.lehmann at jouy.inra.fr]
> > Envoyé : mercredi 16 avril 2008 18:12
> > À : BOURGEOIS Loic
> > Cc : insight-users at itk.org
> > Objet : Re: [Insight-users] 3D separation of touching grains
> >
> >
> >
> > Hi,
> >
> > Le 16 avr. 08 à 17:14, BOURGEOIS Loic a écrit :
> >
> > > Hello everybody,
> > >
> > > I have difficulties to find a method to separate touching grains in
> > > ITK. Has someone an answer to my issue?
> > > Thanks to your help.
> > >
> >
> > A classical way to do that is to compute an distance map inside your
> > objects, and to perform a watershed transform on that distance map.
> > You can build the distance map with
> > itk::SignedMaurerDistanceMapImageFilter for example, and the watershed
> > with MorphologicalWatershedImageFilter.
> >
> > Please look at the example in
> http://insight-journal.org/InsightJournalManager/view_reviews.php?back=search.php%3Ftexte%3Dwatershed&journalid=4&pubid=92
> >  , page 20.
> >
> > Regards,
> >
> > Gaëtan
> >
> > --
> > Gaëtan Lehmann
> > Biologie du Développement et de la Reproduction
> > INRA de Jouy-en-Josas (France)
> > tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> > http://voxel.jouy.inra.fr  http://www.mandriva.org
> > http://www.itk.org  http://www.clavier-dvorak.org
> > __________________________
> >
> > Ce message (et toutes ses pièces jointes éventuelles) est confidentiel et
> établi à l'intention exclusive de ses destinataires. Toute utilisation de ce
> message non conforme à sa destination, toute diffusion ou toute publication,
> totale ou partielle, est interdite, sauf autorisation expresse. L'IFP
> décline toute responsabilité au titre de ce message.
> >
> > This message and any attachments (the message) are confidential and
> intended solely for the addressees. Any unauthorised use or dissemination is
> prohibited. IFP should not be liable for this message.
> >
> > Visitez notre site Web / Visit our web site : http://www.ifp.fr
> > __________________________
> >
>
>  --
>  Gaëtan Lehmann
>  Biologie du Développement et de la Reproduction
>  INRA de Jouy-en-Josas (France)
>  tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>  http://voxel.jouy.inra.fr  http://www.mandriva.org
>  http://www.itk.org  http://www.clavier-dvorak.org
>
>
> _______________________________________________
>  Insight-users mailing list
>  Insight-users at itk.org
>  http://www.itk.org/mailman/listinfo/insight-users
>
>