Re: dicom support

[Andy Buckle]

Thanks, I did mean DS. My implementation gets that right now. (I hope
to have something to share in a few days).

Another request for help:
I hope to implement dicomwrite too. I am not sure there is an easy and
sane way to get back from a name (eg PatientsName) to a tag (eg
0x10,0x10). Any ideas appreciated. I am not too hot at c++, so I may
be missing something in the GDCM doxygen.

Thanks,

Andy

BTW, I switched from the offis dcmdump to the gdcm gdcmdump recently.
It seems to be much faster.

On Sun, May 2, 2010 at 8:12 PM, Judd Storrs <address@hidden> wrote:
> On Sun, May 2, 2010 at 6:52 AM, Andy Buckle <address@hidden> wrote:
>> Also, there are these odd dicom variable representations that store
>> numbers in strings. I don't know if matlab gives them as char arrays,
>> or parses them to arrays of doubles.
>
> I think you are referring to the DS (Decimal String) value
> representation? I did some tests using this DICOM file:
> http://barre.nom.fr/medical/samples/files/MR-MONO2-16-head.gz
>
> Here, I just verify that SpatialResolution is DS using dcmdump from dcmtk:
>
> $ dcmdump +P SpatialResolution MR-MONO2-16-head
> (0018,1050) DS [1.145833\0.859375]                      #  18, 2
> SpatialResolution
> $ grep "1\.145833.0\.859375" MR-MONO2-16-head
> Binary file MR-MONO2-16-head matches
>
> In matlab SpatialResolution is loaded as a 2x1 double vector:
>
>>> dcm = dicominfo('MR-MONO2-16-head') ;
>>> dcm.SpatialResolution
>
> ans =
>
>    1.1458
>    0.8594
>
>>> class(dcm.SpatialResolution)
>
> ans =
>
> double
>
>
>
> --judd
>



-- 
/* andy buckle */