[ITK-users] [ITK] Loading multiframe dicom with itk

[Zein Salah]

Hi Matt,

thanks for the response!

I tested with itk 4.10 and have the following remarks/questions:

1. First, the behavior is different from itk 3.14 and 3.20, but I
think the reader still has bugs.

2. for some data sets, I got a negative z spacing, what does this mean?

3. I have a multiframe-dicom in which the image patient positions
(0020,0032) are stored
individually for each frame under the PerFrame Functional Group
Sequence. Are these positions
taken into account when reconstructing the 3D volume from the block of
data is read?
I am asking because the positions of the slices are going in the
negative z-direction,
which means that the slices are stored in the oopsite order, i.e., the
farthest slice in z
is stored first. Does the reader respect this?

thanks,
Zein



2016-10-19 15:40 GMT+02:00 Matt McCormick <matt.mccormick at kitware.com>:
> Hi Zein,
>
> There has been fixes for this issue in later versions of ITK via both
> ITK's handling of the spacing and the DICOM library used to parse the
> files. It is worthwhile to try with a recent version of ITK to see if
> the issue has been addressed.
>
> HTH,
> Matt
>
> On Wed, Oct 19, 2016 at 6:55 AM, Zein Salah <zeinsalah at gmail.com> wrote:
>> Hi,
>>
>> to organizational reasons, I must use an old release of itk (3.14).
>>
>> I am trying to load multiframe dicom images with itk reader. I noticed
>>  a strange behavior in the itk::ImageFileReader.
>> when I load a multiframe dicom image with the reader, the output of the reader
>> has a correct z-spacing. But the x- and y- spacings are always 1.0. I
>> also could reproduce this using itk 3.20.1.
>>
>> I also noticed that at lease the orientation is wrong.
>>
>> a short search led me to this:
>>
>> http://www.na-mic.org/Bug/view.php?id=725
>>
>> is this aknow issue in itk? has it been fiyed in later versions?
>>
>> Thanks,
>> Zein
>> _______________________________________________
>> Community mailing list
>> Community at itk.org
>> http://public.kitware.com/mailman/listinfo/community