Hi Justin,
Because ImageJ allows a fair bit of freedom in determining the number
of pixels in the canvas you are drawing to, and does not always (and
variably with versions) handle multi-volume / multi-slice data coming
from different source formats as one might expect, I would say it
sounds like at one or another step an image has been defined to be too
big, so the data from four slices got saved into the space allocated
for one.
check with fslhd or fslinfo that your nifti file indeed has 4 slices,
and that the dimensions match your originals
check the header information for the source images make sense within
ImageJ (Image menutab, Properties and Image Information menuitems) and
are consistent with what you are creating at the .hdr stage (there is
an ImageJ nifti writer that should allow you a direct conversion).
ensure that your source images aren't in a montage type format that
ImageJ is clever enough to cut up prior to display but not to writing.
Remember that when you've got that sorted, and importantly if you are
planning to use real nifti functionality downstream... ImageJ's nifti
writer does NOT support the qform and sform information well, and if
you are using .hdr images as an intermediate, the required information
would generally be lost. So you should look to create valid
orientation information for the nifti images (fairly easy if you have
a valid image to copy from, a bit more difficult if not). You have
this when you have the directions labels (LRAPIS) are displayed in
fslview. Related to this, you will have to verify scan per scan that
the slice and left to right ordering of the images in nifti are what
you intended (using ImageJ for DICOM to nifti I've had cases where for
a given sequence a few subjects (out of almost any 30) ended up top to
bottom or left-right flipped, relative to the rest when viewed in
fslview but had the same orientation in ImageJ),
hope it helps
Paul S
>
> ------------------------------
>
> Date: Mon, 12 Aug 2013 16:11:14 +0100
> From: SUBSCRIBE FSL Justin <[log in to unmask]>
> Subject: Converting to NIFTI using ImageJ
>
> I am trying to convert .spr/.sdt files in ImageJ to .hdr to be able to
> convert files in FSL to NIFTI. It appears that my functional data is
> fine. However when I view my anatomical data (only 4 slice coverage), FSL
> displays all 4 slices at the same time ie. first in top left corner, then
> in top right and bottom left and bottom right. Can anyone offer any tips
> on working in ImageJ or FSL to deal with this ie. have FSL display one
> slice at a time so that it will properly do the registration in my later
> steps?
>
> ------------------------------
|