Dear SPMers...
I was following this discussion.
I have problem with the bounding box, too.
I'm using spm12 and I would like to individualize an atlas to each of my subject.
So I have the atlas and I would like to normalize it on my structural space.
In order to evaluate the bounding box values, i'm using the following code:
V=spm_vol('structural.nii');
[BB,vx]=spm_get_bbox(V,'fv');
After the normalization the structural and the normalized atlas have different origin.So after the normalization I tried to cor and reslice the atlas on the structural.
At the end of this process,If I overlap the structrural and the individualized ATLAS, it seems not to match completly.
Why?
Do you have suggestion?
Thank u very much
________________________________________
De : SPM (Statistical Parametric Mapping) [[log in to unmask]] de la part de H. Nebl [[log in to unmask]]
Date d'envoi : jeudi 12 septembre 2013 19:14
À : [log in to unmask]
Objet : Re: [SPM] Normalization ( bounding box)
It is no good idea at all to use several normalisation steps. The default values of the bounding box should be sufficient in general (although you might want to add a few voxels at the front and at the back or at the sides in case of a larger smoothing kernel). If you run into problems as described this is typically due to failure of the normalisation or segmentation as such and not due to the bounding box.
In this case make sure that your images are oriented roughly along AC-PC, similar to the T1 template or the tissue probabiliy maps. Otherwise correct displacements of the origin and tilts manually with the "Display" function.
If you try to normalise on images that are NOT in MNI space you have to adjust the bounding box yourself.
|