> I am trying to use spm5b, but facing some problems.
> 1. When I use the button of DICOM Import to convert the dicom files I
> got from our GE 1.5T MR scanner, I can't find any converted file. And I
> found some error messages in the metlab windows. It said that my Dicom
> files were not dicom files. Why? I can convert these files to analyze7
> format using MRIcro and also using the toolbox of dicom in spm2. Why can't
> spm5 recognize these files?
Is the scanner a GE Twin+excite ? If so, then the latest updates for SPM5b
may solve the problem. The reason for SPM not recognising the data as DICOM
is that it starts with an 8/0 tag, which is not in the DICOM dictionary that
I created a couple of years ago. Maybe there is some new variant of DICOM
that includes this new tag.
> 2. Something interesting was that spm5 could
> recognize the files converted by the dicom toolbox of spm2. So I can
> pre-process the files using the button of realign and normalize.
SPM2 may have been able to convert these files because it didn't check to make
sure the files actually were DICOM. Some people were trying to convert
non-DICOM files with the DICOM toolbox, and wondering why it didn't work.
This is why the check was introduced. I don't know if anyone else has a
better suggestion for recognising (truncated) DICOM files, other than
checking to see if the first few bytes are a valid tag.
> After the
> realignment, I got a img file with the same size(32 MB) to the converted
> img file from the spm2. But after the normalization, the img file was only
> 1 MB. Is that right?
The size of the files depends on the bounding box and voxel sizes specified
for the spatially normalised data.
> 3. I wonder if anyone can give me a detailed steps of
> how to process the data, including how to realign, normalize, and 1st
> specify, especially the step of 1st specify, since spm5 is so different
> from spm99.
The procedure should be pretty similar to that for SPM99. Providing you
understand what you are doing and the reasons for doing them, then the steps
in SPM5 should be very similar.
Best regards,
-John
|