Yet Another Imager Pixel Spacing vs Pixel Spacing
Mathieu Malaterre
Spacing and Imager Pixel Spacing were present in the DICOM file, but
with different value.
In my original post, I am referring to a zip which could be found on
dclunie.com website but was updated in-place, so obviously the old
image is not available anymore.
I am copy/pasting the discussion on wg6:
-------- Original Message --------
Subject: Re: Imager Pixel Spacing vs Pixel Spacing
Date: Mon, 06 Apr 2009 11:49:31 -0400
From: David Clunie <>
Mathieu Malaterre wrote:
> I have been trying to understand CP-586 in the following two cases:
>
> On the one hand:
> - DISCIMG/IMAGES/CRIMAGE taken from
> http://dclunie.com/images/pixelspacingtestimages.zip
>
> And on the other hand:
> - http://gdcm.sourceforge.net/thingies/cr_pixelspacing.dcm
>
> If I understand correctly the CP, one is required to use Pixel Spacing
> for measurement ('true size' print) instead of Imager Pixel Spacing,
> since the two attributes are present and Pixel Spacing is different from
> Imager Pixel Spacing.
>
> If this is correct, then the test data DISCIMG/IMAGES/CRIMAGE is
> incorrect. If this is incorrect (ie. I need to use Imager Pixel
> Spacing), then the display of cr_pixelspacing.dcm for measurement will
> be incorrect.
>
> Could someone please let me know what am I missing here? I could not
> find any information in any header that would allow me to differentiate
> those.
>
> Thank you for your time,
>
Hi Mathieu
The images in pixelspacingtestimages.zip were created before CP 586
(and subsequent related CPs) in order to assess the extent of the
problem
of the presence of pixel spacing attributes with different values.
I have updated the test images to reflect the CP 586 introduction of
the calibration macro, and they now say that Pixel Spacing is being
used instead of Imager Pixel spacing, rather than it is "bad"; they
also say that the user should be informed about what the measured size
"means". See:
http://dclunie.com/images/pixelspacingtestimages.zip
and
http://www.dclunie.com/images/pixelspacingtestimages_screenshot_calibrated.png
That said, if Pixel Spacing and Imager Pixel Spacing are both present,
then the recipient has a dilemma ... which to use and which is
"correct" ?
Given CP 586 etc., I suggest that the logic for a display system
interpreting these attributes for a projection radiograph should be:
- if Pixel Spacing is present and Imager Pixel Spacing is present and
they have the same value then the used should be informed that the
measurements are at the detector plane
- if Pixel Spacing is present and Imager Pixel Spacing is present and
they have different values, then the measurements should use Pixel
Spacing and the user should be informed these are "calibrated"
(in some unknown manner if Pixel Spacing Calibration Type and/or
Pixel Spacing Calibration Description are absent)
- if only Pixel Spacing is present (e.g., in CR IOD), then Pixel
Spacing should be used, but the user should be informed that
what it means is unknown (reiterating that this is only for
projection radiographs - for 3D stuff like CT and MR, that is
the only attribute that should be used and present).
Note that in IHE Mammo profile compliant displays, the value of
Imager Pixel Spacing is required to be corrected by Estimated
Radiographic Magnification Factor and the user informed of that,
hence the additional test case in the revised image set.
David
David Clunie
I have updated these today to include an additional test, and
that is when Pixel Spacing is absent and Imager Pixel Spacing
is present.
This detects systems that just ignore Imager Pixel Spacing
completely (as opposed to always override Imager Pixel Spacing
with Pixel Spacing).
I also added a README file to describe the background behind
these images and what is expected.
David
manu r
Will there be any case where we need to use imager pixel spacing tag and not use pixel spacing tag, even though both tags are present in the image?
Regards,
R Manu