Hi Nick,

The Freesurfer version I use has build stamp freesurfer-Darwin-tiger-i686-stable-pub-v4.0.1

After conversion, read_asc.m doesn't work - seemingly due to a dimension problem. In the code I see that 5 elements are expected for faces and indicies, whereas only 4 elements are provided in the .asc file. I changed lines 45-46 into

f = fscanf(fp, '%d', S(3)*4); f = reshape(f, [4 S(3)])';

Using patch in Matlab with v(:,1:3) as vertices and f(:,[1 3 2]) as faces, I get an image of the hippocampus surface.

The surface looks good(!) but is a bit rough and 'spiky'. Is this because the segmentation procedure does not employ a smoothness constraint?

The aim of all this is to prepare hippocampal volumes for deformation analysis (two groups, two time points). Is the following line of processing appropriate:

Topology correction: mri_topologycorrection -label "hippocampus" subject/mri/norm.mgz subject/mri/aseg.mgz aseg_corrected

Tesselate surface: mri_tessellate

Smooth surface: mris_smooth

Convert to ascii for use with Matlab: mris_convert

Cheers! Kim.

2007/11/16, Nick Schmansky nicks@nmr.mgh.harvard.edu:

Kim,

Which version of freesurfer is being used?

An alternative approach is to convert the surface to ascii:

mris_convert surface_file surface_file.asc

where the .asc extension is an output ascii file listing the vertices and faces.

Nick

On Thu, 2007-11-15 at 16:18 +0100, Kim Mouridsen wrote:

...

Hi,

We have segmented hippocampus using Freesurfer, and want to generate a surface which can be read into matlab.

We use mri_tesselate and mris_smooth. However, the resulting file can not be read into matlab using read_surf. The error we get is

??? Undefined function or variable "vertex_coords".

Error in ==> read_surf at 77 vertex_coords = reshape(vertex_coords, 3, vnum)' ;

It seems the 'magic' number is 16777213, whereas only 16777214 and 16777215 are supported in read_surf.m

How can we read the result of mris_smooth into matlab?

Thanks!

Kim. _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

Hi Bruce,

Thanks, that would be perfect!

I tried increasing the number of smoothing iterations in mris_smooth, and this improves the result, but this is probably not the best way to do it.

We would be very happy if you could send us a beta version of your smoothing tools.

Cheers, Kim.

2007/11/16, Bruce Fischl fischl@nmr.mgh.harvard.edu:

Hi Kim,

we do have some tools for smoothing the hippocampal surfaces, but we haven't distributed them. I could try to get you a beta version if you like.

cheers, Bruce On Fri, 16 Nov 2007, Kim Mouridsen wrote:

...

Hi Nick,

The Freesurfer version I use has build stamp freesurfer-Darwin-tiger-i686-stable-pub-v4.0.1

After conversion, read_asc.m doesn't work - seemingly due to a dimension problem. In the code I see that 5 elements are expected for faces and indicies, whereas only 4 elements are provided in the .asc file. I changed lines 45-46 into

f = fscanf(fp, '%d', S(3)*4); f = reshape(f, [4 S(3)])';

Using patch in Matlab with v(:,1:3) as vertices and f(:,[1 3 2]) as faces, I get an image of the hippocampus surface.

The surface looks good(!) but is a bit rough and 'spiky'. Is this because the segmentation procedure does not employ a smoothness constraint?

The aim of all this is to prepare hippocampal volumes for deformation analysis (two groups, two time points). Is the following line of processing appropriate:

Topology correction: mri_topologycorrection -label "hippocampus" subject/mri/norm.mgz subject/mri/aseg.mgz aseg_corrected

Tesselate surface: mri_tessellate

Smooth surface: mris_smooth

Convert to ascii for use with Matlab: mris_convert

Cheers! Kim.

2007/11/16, Nick Schmansky nicks@nmr.mgh.harvard.edu:

...

Kim,

Which version of freesurfer is being used?

An alternative approach is to convert the surface to ascii:

mris_convert surface_file surface_file.asc

where the .asc extension is an output ascii file listing the vertices and faces.

Nick

On Thu, 2007-11-15 at 16:18 +0100, Kim Mouridsen wrote:

...

Hi,

We have segmented hippocampus using Freesurfer, and want to generate a surface which can be read into matlab.

We use mri_tesselate and mris_smooth. However, the resulting file can not be read into matlab using read_surf. The error we get is

??? Undefined function or variable "vertex_coords".

Error in ==> read_surf at 77 vertex_coords = reshape(vertex_coords, 3, vnum)' ;

It seems the 'magic' number is 16777213, whereas only 16777214 and 16777215 are supported in read_surf.m

How can we read the result of mris_smooth into matlab?

Thanks!

Kim. _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

There is a freely available MATLAB codes for smoothing arbitrary surface objects using spherical harmonics. It's a different technique. But this technique will provide surface-to-surface alignment automatically without additional computation. See

http://www.stat.wisc.edu/~mchung/softwares/weighted-SPHARM/weighted-SPHARM.h...

You may have to change code a bit since "mesh connectivity" will be different from one surface to another.