[caret-users] Resampling a surface

[David Van Essen]

Andrew,

To amplify on what Donna said about the Caret process.

The closest that Caret software and atlases come to providing uniform  
resampling is through the use of our 'standard-mesh', 73,730-node  
spherical mesh:

http://sumsdb.wustl.edu/sums/archivelist.do?archive_id=6362810&archive_name=Human.PALS_B12.SPHERE.73730.coord
http://sumsdb.wustl.edu/sums/archivelist.do?archive_id=6362816&archive_name=Human.sphere_6.RIGHT_HEM.73730.topo

This sphere provides relatively uniform node spacing on the sphere.   
(It's not perfectly uniform for technical reasons that I can elaborate  
on if you care.)

If your surface of interest includes a spherical configuration that is  
low-distortion relative to the 'fiducial' surface (by Caret's multi- 
resolution morphing or by the distinctly different algorithms that  
FreeSurfer and other apps use), you can obtain standard-mesh  
resampling of each of these:

"-deformation-map-create" creates a deformation map from one sphere to  
another (this assumes that spheres are co-registered).  The output of  
the command is a deformation map that maps from one sphere (e.g., a  
FreeSurfer sphere) to another sphere (e.g., the PALS standard-mesh  
sphere).  Once the deformation map is created, the command "- 
deformation-map-apply" is used to deform data files, such as  
coordinate files.

Because there are inevitably large distortions in mapping to a sphere  
(especially in and near the Sylvian fissure and temporal pole),  
surfaces resampled by this strategy will have nonuniform node spacing  
of their 3D (fiducial) surfaces.

David

On Aug 7, 2008, at 7:57 AM, Donna Dierker wrote:

> Hi Andrew,
>
> I'm not aware of a Caret feature that ensures uniform resampling, but
> the caret_command -deformation-map-apply and -deformation-map-create
> features do support non-uniform resampling.  Simplify surface will  
> just
> reduce the number of nodes; it doesn't ensure uniform spacing.
>
> Have a look at AFNI/SUMA's IsoSurface command line utility.  The  
> usage says:
>
>>  NOTE:
>>  The code for the heart of this program is a translation of:
>>  Thomas Lewiner's C++ implementation of the algorithm in:
>>  Efficient Implementation of Marching Cubes´ Cases with Topological
>> Guarantees
>>  by Thomas Lewiner, Hélio Lopes, Antônio Wilson Vieira and Geovan  
>> Tavares
>>  in Journal of Graphics Tools.
>>  http://www-sop.inria.fr/prisme/personnel/Thomas.Lewiner/JGT.pdf
> The name suggests uniform node spacing, although I could be
> misinterpreting it.
>
> The entire suite of AFNI/SUMA goodness awaits you at:
>
> http://afni.nimh.nih.gov/afni/
>
> SUMA can read both Freesurfer and Caret formats, and probably the new
> GIfTI surface standard as well.
>
> Donna
>
> On 08/07/2008 04:03 AM, Andrew Reid wrote:
>> Hi all,
>>
>> I need to get a uniform resampling of a cortical surface mesh.  
>> Caret has
>> a "Simplify surface" command, however this appears to perform a
>> selective resampling (i.e., preferentially removing nodes which carry
>> less information), so this is not ideal for us. Is there another  
>> way to
>> do this with Caret? I think that the deformation map process can also
>> result in a resampled surface, but I'm not sure of the best way to  
>> do this.
>>
>> Cheers,
>> Andrew
>>
>>
>
> _______________________________________________
> caret-users mailing list
> caret-users at brainvis.wustl.edu
> http://brainvis.wustl.edu/mailman/listinfo/caret-users