Hi,
The first image in the 4D file is the sum of all the consecutive images
after they have been boundary corrected (using first_utils). In the
uncorrected image, the boundary voxels are equal to 100+ the structure
label. When first_utils is run, some boundary voxels are threshold off,
and the remaining are set the the structures label. I think your
differences are due to the fact that some voxels have not been the
thresholded nor have the other voxel been set to the proper structure
label.
Hope this helps,
Cheers,
Brian
> Hi,
>
> I changed the last lines of the run_first_all script. Now, I have the
> segmented
> images of all the structures seperately as Mark said before. I merged all
> of
> them together and created a 4*D file. I want to have an image the same as
> the 4*D file that the run_first_all command generates. As you know,
> run_first_all command generates a 4*D file with 18 images in t direction.
> The
> first one is the average image of 17 structures and the remailning 17
> images
> are the structures themselves. I compared mine to an image that I had
> previously generated by successfully running the run_first_all command. It
> seems that they are not the same. Is there any commant? Besides, what is
> the difference beween image1_BrStem_corr.nii.gz and
> image1_BrStem_first.nii.gz? I use image1_BrStem_first.nii.gz to generate
> my
> 4*D image. I use the following commands:
> -fslmerge -t image1-4*D all the structures
> -fslmaths image1-4*D -Tmean -mul 17 image0 -odt int
> -fslmerge -t image2-4*D image0 image1-4*D
> image0 (first image in image2-4*D) should be the same as the result of
> run_first_all. But they are different.
>
> Siamak
>
>
|