Hi Flavio,
Something looks very wrong to me with your first image, if the voxel
dimensions really were 1x1x1 then there would need to be a lot more
than 50 voxels in the z-direction (anterior-posterior) to cover the
brain. I think the true voxel dimensions must be non-isotropic, more
like 1x1x2. I expect the coregistration fails because this image and
the second (which looks a lot more like 1x1x1) are initially quite
badly mismatched in z-scale.
I think the best thing to do would be to check your preprocessing
(e.g. dicom conversion) to try to ensure that the voxel dimensions of
the first image are correct. An alternative might be to give the
coregistration an initial set of parameters including a factor of two
(or half depending on direction) stretch in the z-direction.
Best,
Ged.
Flavio wrote:
> SPMers,
>
>
>
> I’m trying getting SPM multi-spectral segmentation through SPM 2. I’ve
> worked with two MR acquisition of same subject: T1-weighted and IR.
>
>
>
> As showed figures bellow, the normalization seems to work well. But,
> when I try computing both images coregistration, I haven’t get success.
> The coregistration and segmentation results are also showed bellow.
>
>
>
> I would appreciate some suggestions as to getting better results. Any
> doubt will be welcome.
>
>
>
> Thanks,
>
> Flávio.
>
>
>
>
>
|