[Insight-users] Problem reading 4D Dicom datasets

[Antoine DUCHAMPS]

Hi Mathieu,

gdcmtar works fine and I can recover the correct volume size (128 x 128
x 41) after spliting a 3D file into a set of 2D images.

However, back to the original problem, now I obtain an image (128 x 128
x 2367 x 1) which is still a 3D image. Now the problen seems to be the
ordering of the 4th dimension as you commented previously. You had
proposed me to use the instance number. My question is: at which level
do I have to specify this? Is this a function of the ImageSeriesReader
or a function of the gdcmImageIO? 

thanks!

antoine

PS : I don't know if it's important for the frame ordering but I get the
following errors too: 

Error:
In /home/duchamps/lib/itk-3.10.2-src/Utilities/gdcm/src/gdcmSerieHelper.cxx, line 663, function bool gdcm::SerieHelper::ImagePositionPatientOrdering(gdcm::FileList*)
File: /home/duchamps/Data/dti_adulte/dicom_untar/2940 Distance: 68.5837
position is not unique




Le jeudi 23 avril 2009 à 10:04 +0200, Antoine DUCHAMPS a écrit :
> Your cristal ball is right ! :) 
> I'll try with gdcmtar and I'll go back to you.
> Thanks a lot!
> 
> antoine
> 
> 
> 
> 
> 
> Le jeudi 23 avril 2009 à 09:53 +0200, Mathieu Malaterre a écrit :
> > On Thu, Apr 23, 2009 at 9:45 AM, Antoine DUCHAMPS
> > <antoine.duchamps at gmail.com> wrote:
> > > This is the output of gdcminfo* for the first volume of the sequence.
> > >
> > > Hope this helps,
> > >
> > > antoine.
> > >
> > > * I've used gdcminfo of gdcm 2.0.10 since I didn't find it for gdcm
> > > 1.2.x
> > >
> > >
> > > MediaStorage is 1.2.840.10008.5.1.4.1.1.4 [MR Image Storage]
> > > TransferSyntax is 1.2.840.10008.1.2.1 [Explicit VR Little Endian]
> > > NumberOfDimensions: 2
> > > Dimensions: (896,896,1)
> > 
> > So indeed your image is 896x896, no doubt about it.
> > 
> > My cristal ball is telling me that you are dealing with SIEMENS MOSAIC
> > file. Is this correct ? If this is correct simply use the gdcmtar
> > (same location as gdcminfo you just used) to split your file into a
> > valid multiframe DICOM file.
> > 
> > http://apps.sourceforge.net/mediawiki/gdcm/index.php?title=Gdcmtar
> > 
> > It wil properly discard any black frame out of the final multiframe DICOM file.
> > 
> > HTH