[Insight-users] Problem reading 4D Dicom datasets

[Mathieu Malaterre]

On Tue, Apr 21, 2009 at 11:07 AM, Antoine DUCHAMPS
<antoine.duchamps at gmail.com> wrote:
> Hi all,
>
> I'm trying to read a 4D dataset (DWI) in Dicom format by using a
> modified version of DicomSeriesReadImageWrite2. I just set Dimension = 4
> and commented the line "nameGenerator->SetUseSeriesDetails( true );"
> My dataset is organized as a set of 3D files (one volume for each
> gradient orientation).
>
> The problem I have is that I should obtain an image [128 128 41 62] and
> I get an image [896, 896, 62, 1]. I've printed out the filenames created
> by nameGenerator and they seem to be correct (they are the file names of
> all 3D files)
>
> I've searched the list but I haven't found the solution :(. Can anybody
> help?

Ordering DWI data is IMHO a hard task. DICOM does not defines AFAIK
the ordering for those dataset. So what you end up with a set of 3D
volumes but you have no clue which one comes first, which one comes
last...
To make things worse, a well known private vendor does not store
exactly(*) the same Image Orientation (Patient) within the same DWI
volume, which makes writing conditions even more difficult.