I realize this is a very old thread, but in case anyone finds this via a web search.... when you are specifying a gaussian you are describing the sigma - voxels further away will still have a significant contribution. To capture this, fslmaths will make a kernel that is 6-sigma wide (so a diameter of 6, and radius of 3). Therefore, from first principles one would expect a "-kernel gauss 10" to create a radius 30 spherical kernel with 27 times the volume of "-kernel sphere 10". In practice, fslmaths generates a cube with wxhxd of 60x60x60mm, so 52 times the area of the sphere. For this reason, usage of "-kernel gauss" has tremendous computational requirements. Fortunately, the Gaussian smooth is mathematically separable, so we can get the identical result by solving each dimension independently. Therefore, one should use the fslmaths function "-s <sigma>" instead of the "-kernel gauss <sigma> -fmean".
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|