[Insight-users] Pixel Spacing / DICOM Series (was Re: InsightToolkit-3.12.0 problem reading Structure DCM)

[Mathieu Malaterre]

Maximilien,

On Tue, Sep 15, 2009 at 1:46 PM, Maximilien Renard <iixamaxii at gmail.com> wrote:
> Hi,
>
> Sorry for my late answer.
>
> So I've checked the dicom tags of my image using
> DicomImageReadPrintTags and I've highlighted these lines :
>
> (0018|0050) Slice Thickness = 0.75
> ...
> (0020|000e) Series Instance UID = 1.3.12.2.1107.5.1.4.24514.4.0.2018232119198399
> ...
> (0028|0030) Pixel Spacing = 0.41015625\0.41015625
>
> So I've read a little bit on the internet and it seems that Slice
> Thickness is not what I'm searching for.
>
> The good way seems to be these tags :
>
> (0020|0032) Image Position (Patient) = -101.79492\-211.79492\-1107.5
> (0020|0037) Image Orientation (Patient) = 1\0\0\0\1\0
>
> and use them in order to :
>
> 1) compute the normal of the planes
> 2) use the Image Position to compute the position of the current slice
> on the axis (and the difference between the position of two adjacent
> slices gives the spacing between them).
>
> ( http://newsgroups.derkeiler.com/Archive/Comp/comp.protocols.dicom/2006-01/msg00055.html
> )
>
> Since I am kind of a newbie, please forgive me if I asking
> stupidities, here are my questions :
>
> 1) How come that despite the fact that there is this Series ID, my
> images are not a serie ?

I am not sure I understand your question. Do you understand that DICOM
Series does *not* mean homogenous volume (i.e constant z-spacing)

> 2) Will the method to compute the Z-Spacing that I describe above work
> for all kind of CT Images ?

This is internally done you should not even need to know how it work.

> 3) Doesn't ITK have a way to determine this spacing ?

Yes it shoudl work by default.

-- 
Mathieu