In that case, GE must be setting the fields in the DICOMs in each set of
999 DICOMs in some manner such that even after you put them all into a
single folder, neither dcm2nii or MRIconvert are able to decipher how to
group the dicoms appropriately into volumes and time points.
i.e., Your problem is with the initial conversion, not with fslmerge.
cheers,
-MH
On Tue, 2012-05-29 at 21:24 +0100, Marta Penedo wrote:
> I said this because I tried it both ways. But yes, I already tried
> that.
>
>
> The name's change was only another attempt.
>
>
>
>
> Marta
>
> 2012/5/29 Michael Harms <[log in to unmask]>
> So, you had a single folder with the complete set of 3640
> DICOMs from
> the run, and ran dcm2nii on that folder? Because you
> initially wrote
> that you ran dcm2nii and MRIconvert on the incomplete folders.
>
> As far as I know, the actual filename of the DICOMs are
> irrelevant, as
> the NIFTI is assembled based on the fields in the DICOMs.
>
> cheers,
> -MH
>
> On Tue, 2012-05-29 at 20:49 +0100, Marta Penedo wrote:
> > I put all dicom images in a unique folder and changed their
> filename
> > to the same filename (in sequence), and then I tried to
> convert images
> > of this folder, but the convertion does 4 nii files such as
> the images
> > are separated.
> >
> >
> > Is that what you asked me?
> >
> > 2012/5/29 Michael Harms <[log in to unmask]>
> > Hi Marta,
> > Did you try moving all the DICOMs from a given run
> into a
> > single
> > directory PRIOR to converting to NIFTI? You can't
> convert to
> > NIFTI
> > using an incomplete set of DICOMs and expect the
> converter to
> > necessarily handle it cleanly.
> >
> > cheers,
> > -MH
> >
> >
> > On Tue, 2012-05-29 at 19:59 +0100, Marta Penedo
> wrote:
> > > Yes, I see this in mricron too.
> > >
> > >
> > > In this case, the number of slices is 26 and 140
> volumes, of
> > a total
> > > of 3640 images, but this information is not in
> images. I
> > think that
> > > acquisition was not well done, but the files have
> been used.
> > >
> > >
> > >
> > > 2012/5/29 Jesper Andersson <[log in to unmask]>
> > > Hi again Marta,
> > >
> > > >
> > > >
> > > > First nii file: dim1=64, dim2=64,
> dim3=973
> > > > Second nii file: dim1=64, dim2=64,
> dim3=999
> > > > Third nii file: dim1=64, dim2=64,
> dim3=999
> > > > Fourth nii file: dim1=64, dim2=64,
> dim3=669
> > > >
> > > >
> > > > In all cases, the matrix are 64x64 and
> the dim3
> > > > corresponding at number of images in
> each folder
> > and nii
> > > > file appears such a 3D file.
> > >
> > >
> > > I am a little confused by this. Usually
> dim1 and
> > dim2 are the
> > > matrix size (which seems to be true here
> too) and
> > dim3 are
> > > number of slices (which it is clearly not
> here).
> > What do you
> > > see if you open one of those nifti files
> in fslview?
> > Do you
> > > get a bunch of brains stacked on top of
> each other?
> > >
> > >
> > > Jesper
> > >
> > >
> > >
> > > >
> > > > 2012/5/29 Jesper Andersson
> <[log in to unmask]>
> > > > Chose one of the cases that
> doesn't work
> > and type
> > > >
> > > >
> > > > fslhd my_nifti | grep dim
> > > >
> > > >
> > > > for each of your nifti files and
> check the
> > values of
> > > > dim1 to dim3
> > > >
> > > >
> > > > Jesper
> > > >
> > > > On 29 May 2012, at 13:05, Marta
> Penedo
> > wrote:
> > > >
> > > > > The command that I used was:
> fslmerge -t
> > FFfmri7
> > > > > F*
> > > > >
> > > > >
> > > > > I think that the command is
> correctly.
> > The problem
> > > > > is that images are grouped, in
> same
> > cases, in 4
> > > > > folders (one with 976, 2
> folders with
> > 999 images
> > > > > and other folder with 627) and
> in other
> > cases in 2
> > > > > folders (one with 984 and
> other with
> > 617). The
> > > > > different information between
> images of
> > each
> > > > > folders are in
> SeriesInstanceUID, Series
> > Number,
> > > > > Instance Number, Image
> Position Patient
> > and Slice
> > > > > Location.
> > > > >
> > > > >
> > > > > I tried to merge the two
> folders with
> > 999 images,
> > > > > for experience, and the
> warning
> > disappear. But I
> > > > > have to merge all of them.
> > > > >
> > > > >
> > > > > Is There any way to change the
> > information without
> > > > > interfering with images?
> > > > >
> > > > > 2012/5/29 Jesper Andersson
> > <[log in to unmask]>
> > > > > Dear Marta,
> > > > >
> > > > >
> > > > > > I tried to convert
> dicom
> > images,
> > > > > acquired from 1.5T GE
> CVi/NVi,
> > to NIfTI.
> > > > > This device allows
> only acquire
> > 999 images
> > > > > at a time, but the
> complete
> > sequence have
> > > > > 3640 images. So,
> software put
> > each group
> > > > > of images in different
> folders,
> > such as
> > > > > "RTIP_6" and
> "RTIP_26".
> > > > > >
> > > > > > I tried to convert
> and merge
> > all images
> > > > > using dcm2nii and
> MRIConvert,
> > but this
> > > > > softwares shows me one
> file .nii
> > for each
> > > > > folder. Then, I tried
> to merge
> > all
> > > > > files .nii obtained
> using
> > fslmerge but the
> > > > > command shows me this
> warning:
> > > > > >
> > > > > > Error in size-match
> along
> > non-
> > > > > concatenated
> dimension.
> > > > >
> > > > >
> > > > > If you have run
> fslmerge
> > correctly this
> > > > > would imply that you
> have
> > different image
> > > > > sizes for the
> different "sub
> > series",
> > > > > which sounds very
> strange. If
> > that is the
> > > > > case you should look
> into the
> > reason for
> > > > > that.
> > > > >
> > > > > Can you send us the
> exact
> > fslmerge command
> > > > > you used?
> > > > >
> > > > > Jesper
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
> >
>
>
>
|
