Hi Kamran,
I'm not sure I've understood exactly what you've done, but in general,
if you use world/mm coordinates for landmarks then you should be able
to compare distances and positions across different resolution images
(with the caveat that very low resolution might give rather unreliable
results). If you use V=spm_vol('image.img'); then V.mat*[x y z 1]';
will give you the world coordinates for voxel [x y z].
It sounds like your synthetic images might not have correct
voxel-world mappings though... How are you generating them, and what
is it that you are really trying to do?
Best,
Ged.
kamran kazemi wrote:
> Dear SPMers
>
> I have question about normalization that maybe is simple. (so sorry at
> first)
>
> The question is:
> I created two synthetic images with two different resolutions, first one
> with 1*1*1 mm and the second one with 0.5*0.5*0.5mm, and both are on the
> same matrix size (for example 100*100*100 voxels) and also they both
> fill the same percentage in their volume. For example both objects fill
> 50% of the volume and the rest is background in both.
> Now, I normalize a subject with resolution 0.5*0.5*0.5mm to these two
> images using normalization process in SPM. So finally I have two
> normalized images based on the two reference images which have 1*1*1mm
> and 0.5*0.5*0.5mm resolution.
>
> So, the question is that:
> I define a landmark on the reference images and find it in the
> normalized images based on these two reference images. So I have two
> distances d1 and d2 that come from landmarks on two different resolution
> reference images. So is it fair to compare these two lengths directly?
> Or one should use relative length for comparison?
>
> Kind Regards
> kamran
>
> ------------------------------------------------------------------------
> Have a burning question? Go to Yahoo! Answers
> <http://answers.yahoo.com/;_ylc=X3oDMTFvbGNhMGE3BF9TAzM5NjU0NTEwOARfcwMzOTY1NDUxMDMEc2VjA21haWxfdGFnbGluZQRzbGsDbWFpbF90YWcx>
> and get answers from real people who know.
|