[Insight-users] NRRD output issue

[Soubzriquet]

Following up my own question:

I should have tried this before, but if I write out the original image
everything is fine (NRRD0005, and the measurement frame is written out).
 So that answers half of my issue.

So my refined question is this:

I am constructing a new vector image using itk:ComposeImageFilter  and
averaging operations on images taken from an
itk::VectorIndexSelectionCastImageFilter.

All other spatial information is being carried along with the sliced
images, but not the measurement frame.  Can anyone point me in the right
direction to preserve this information?





On Tue, Jul 9, 2013 at 4:16 PM, Soubzriquet <soubzriquet at gmail.com> wrote:

> Hi all!
>
> I have the following issue where I am losing metadata when outputting to a
> NRRD file.
>
>  1) I load a valid NRRD0005 file with a "measurement frame" field into an
> itk::VectorImage
>  2) I filter the image, creating a new vector image
>  3) I write out the new image to a file
>
> The resulting file is in NRRD0004 format, and does not include a
> measurement frame.
>
> I've dumped the entire metadata dictionary and I can see that the
> measurement frame is being read from the file, but it is handled
> differently that other properties like "kinds" and "space",  it's metadata
> object type name is std::vector<std::vector<double>>> whereas the other
> types show up as std::strings.
>
> So it looks like I can special case this field and regenerate it to hack
> around the problem, but I'd  like to see how to do this properly.
>
>
> So I am wondering does ITK 4.3.2 support NRRD0005 at all?  If so, is this
> a known issue or am I just doing something incorrectly?
>
> Any ideas how best to proceed before I dive into the ITK source?
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20130709/7f2500d7/attachment.htm>