[MINC-users] Getting max and min in a sphere centered at each voxel

[Alex Zijdenbos]

I realized there were some issues with the kernel generation script I passed along (center voxel nonzero, among other things), so here is an improved version:

make_mincmorph_kernel.pl (http://cl.ly/3P0i391X0k0e2V1T2846)

It now also allows for ellipsoidal kernels. As examples, the 4 standard mincmorph kernels can be generated as follows:

make_mincmorph_kernel.pl -radius 1 3D06.ker
make_mincmorph_kernel.pl -radius 1.8 3D26.ker
make_mincmorph_kernel.pl -radius3 1 1 0 2D04.ker
make_mincmorph_kernel.pl -radius3 1.8 1.8 0 2D08.ker

-- A

On Thursday, 5 July, 2012 at 8:10 PM, Alex Zijdenbos wrote:  
> Ah - so mincmorph actually does grey level morphology? Somehow I've only ever used it for binary morphology. Nice.  
>  
> However, I do think you meant:
>  
> mincmorph -3D26 -erosion in.mnc min.mnc
> mincmorph -3D26 -dilation in.mnc max.mnc
>  
> Then with this perl script:
>  
>  ​make_mincmorph_kernel.pl (http://cl.ly/2420323b382t2z3p3E3N)
>  
> you can generate spherical kernels of arbitrary size; but yes it will get a bit slow for large ones. I have generated 10x10x10 kernels though and that is all still very doable (you couldn't finish your coffee in the time it takes to run).
>  
> -- A
>  
> On Thursday, 5 July, 2012 at 5:53 PM, Andrew Janke wrote:
> > mincmorph will also do this but only behaves well for small
> > structuring elements (up to 5x5). Beyond that it's a tad slow.
> >  
> > You should be able to get a rough hack at what you want with something
> > like this for a 3x3x3 structuring element to start with:
> >  
> > mincmorph -3D26 -erode in.mnc max.mnc
> > mincmorph -3D26 -dilate in.mnc min.mnc
> >  
> > This should then get what you want across the whole volume in min.mnc
> > and max.mnc.
> >  
> >  
> > a
> >  
> > On 6 July 2012 07:49, Vladimir S. FONOV <vladimir.fonov at gmail.com (mailto:vladimir.fonov at gmail.com)> wrote:
> > > Then you would need to use gray-scale morphology and perform dilation to get
> > > max and erosion for min. I have a tool based on ITK which does that .
> > >  
> > >  
> > > On 12-07-05 04:39 PM, francois hebert wrote:
> > > >  
> > > > Hi,
> > > >  
> > > > you suggestion seems good for one voxel but I would like a filter that
> > > > will
> > > > go through all the voxels.
> > > >  
> > > > Francois Hebert
> > > >  
> > > > 2012/7/5 Vladimir S. FONOV <vladimir.fonov at gmail.com (mailto:vladimir.fonov at gmail.com)>
> > > >  
> > > > > Hello,
> > > > >  
> > > > >  
> > > > > On 12-07-05 03:38 PM, francois hebert wrote:
> > > > >  
> > > > > > I would like to know if there is a minc tool that allow to get the max
> > > > > > and
> > > > > > min inside a sphere of a predefined diameter at each voxel.
> > > > > >  
> > > > >  
> > > > >  
> > > > > i would think make_phantom to create a spherical mask followed by
> > > > > mincstats -min -max
> > > > >  
> > > > > --
> > > > > Best regards,
> > > > >  
> > > > > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> > > > >  
> > > > >  
> > > > > ______________________________**_________________
> > > > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > > >  
> > > > > http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
> > > > _______________________________________________
> > > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > > >  
> > >  
> > >  
> > >  
> > > --
> > > Best regards,
> > >  
> > > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> > >  
> > >  
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > >  
> >  
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >  
> >  
> >  
>  
>