Print

Print
You need to ensure that you have sensible headers for your images.  Make sure voxel sizes are correct, that intensities are between 0 and 1, and that for each subject, the tissue class probabilities don't sum to a value greater than 1.  Also, quite a common mistake I've seen people make is to specify their GM and WM tissue classes together.  These should be entered as two separate sets of data, with the order of the GM being the same as the order of the WM.

Best regards,
-John




On 20 May 2013 16:42, Emilie Tuynder <[log in to unmask]> wrote:
Dear, 

I would like to use Dartel on a set of segmentations which were not created by SPM.  If I use the segmentations directly, the results look really weird ( zoomed in on some voxels, the numbers that are printed in the workspace during the program are also of size E09). I looked at previously rc1 created images, resampled my images to the same voxel size and data type but the results are still bad. The first template template_0 looks good, so I think something goes wrong in the registration. In the manual it specifies that images need to be created that DARTEL can work with it. Although it does not specify what this actually means. Can anybody tell me what information needs to be added so DARTEL can work with my segmentations? 

Thanks in advance, 
Kind regards, 
Emilie