On Wed, 17 Aug 2005 11:11:41 +0100, S.P. Chen <[log in to unmask]>
wrote:
>Hello
>
>I am new to SPM. I plan to use SPM2 to fuse SPECT image and MRI(T1) image
>(the ".analyze" from the same patient) together recently. The attachment
>of "coregister.jpeg" file is the result I got after coregister. Before
>coregister, I reslice(a command in coregister) the MR image and using its
>SPECT image as a template.
In general, the two images don't need to have the same number of slices,
so this step is redundant. I'd recommend not doing it.
>It makes SPECT and MR(T1) image have the same
>number of slice. Then I normalize both SPECT and MR image(after reslice)
>with SPECT and T1 templates that built in SPM2 respectively.
Unless there's something special about SPECT images I don't know about, if
the images come from the same subject, you don't need to spatially
normalize before coregistering. In fact, it's probably a bad idea.
>Then I use MR
>(T1) image as a target and SPECT image as a source to run coregister
>command. I export the result to the "coregister.jpeg" file. There are four
>figures in the result. The lower two figures seem MR and SPECT I used to
>fuse.
I might be looking at it wrong, but the quality of the registration seems
poor. In particular, the sagittal display makes it look like the two
brains are not "pitched" the same amount.
>But I do not know what the upper two figures are.
The upper two figures are joint histograms. It's technical information
having to do with the nature of the algorithm used in coregistration.
Just ignore it unless you have expert knowledge about the algorithm.
>Is this the
>correct result I should get after coregister? I can not find the correct
>fusion image. The upper two figures look like some function curve. May you
>tell me what wrong with it is? Is there still any step I do not progress?
>There is still another question. While I normalize the MR(T1) image,
>should I use T1 template or SPECT template? In this result, I use T1
>template.
When you spatially normalize an image, you should always use a template of
the same modality (T1 ==> T1, PET ==> PET, EPI ==> EPI, etc).
>
>Thank you for your help in advance.
>
|