Dear Rudolph,

Thanks a lot for your reply! Do you know why does it fail on holes/edges? Does the algorithm require each triangle vertex to have 1-ring neighbours?

Also, do you by chance know how to use the -a parameter correctly? I did not observe any difference nor in the results nor in runtime with a call like this:

mris_curvature_stats -a 5 -f surface -m -o noisy_sphere_r10_a5.stats -G --signedPrincipals --writeCurvatureFiles noisy_sphere noisy_sphere_r10

Many thanks and best wishes

Maria

On 30.07.19 18:00, freesurfer-request@nmr.mgh.harvard.edu wrote:

Hi Maria --

If there are holes/edges in the surface `mris_curvature_stats` it will fail.