[Insight-developers] gdcm patient orientation

[Mathieu Malaterre]

And I have the exact same understanding of the DICOM spec (any 
mathematical operation should transform to Secondary Capture). Now 
rereading some post in dicom newsgroups it looks like I can set it to 
Secondary Capture, but still keep original Modality.

http://groups.google.com/group/comp.protocols.dicom/msg/79e4dd33b9de9540

It looks like I either need some other kind of information like number 
of frames: 0028 0008 IS 1 Number of Frames, or -hopefully- leaving 
Modality to CT, will make the GE A/W happy.

2 cents
Mathieu
Ps: of course there is one case, where no mathematical operation is done 
and in this case then yes GDCM should allow user to set whatever SOP 
class they want...

Stephen R. Aylward wrote:
> Hi,
> 
> It would be great to get an expert's opinion on secondary capture 
> images.   My understanding is that after image processing the image is 
> considered a secondary capture image - results of digital subtraction 
> are one of the most commmon examples.   After blurring or edge enhancing 
> an MR image, is it still an MR image?
> 
> Stephen
> 
> 
> 
> 
> 
> Bill Lorensen wrote:
> 
>> That may be because gdcm is still reporting the new series as a 
>> Secondary Capture. I'm not sure. In my local copy I have removed 
>> 0002|0002 and 0008|0016 from gdcmFileHelper. I do not think the new 
>> series should be a secondary capture. Rather it is really a derived 
>> image.
>>
>> Since 0020|0020 does not affect me, you can leave it in. But, saying 
>> the series generated by itk are Secondary Captures is not correct I 
>> believe.
>>
>> Bill
>>
>> At 11:13 AM 2/21/2006, Mathieu Malaterre wrote:
>>
>>> Lorensen, William E (GE, Research) wrote:
>>>
>>>> Mathieu,
>>>> My dicom checker complains:
>>>> (0x0020,0x0020)  CS  4        [L\P] # Patient Orientation
>>>> !!                            ERROR : Patient Position 
>>>> (0x0020,0x0020) must not be present
>>>> !!                                    if (0x0020, 0x0032) and 
>>>> (0x0020, 0x0037) are required
>>>> In Table C.7-9 of the dicom standard:
>>>> Patient direction of rows and columns of the image. Required if 
>>>> imkage does not require Image Orientation (Patient) (0020,0037) and 
>>>> Image Position (Patient) (0020,0032). See C.7.6.1.1.1 fr further 
>>>> explanation.
>>>>
>>>> I think gdcm should not be adding 0020|0020 if the others are 
>>>> present. However, this does not seem to affect processing on the A/W.
>>>
>>>
>>>
>>> If I remove 0020|0020, David Clunie tools report:
>>>
>>> ./dciodvfy Image0075.dcm
>>> ...
>>> Error - Missing attribute Type 2C Conditional 
>>> Element=<PatientOrientation> Module=<GeneralImage>
>>>
>>> Mathieu
>>>
>>> Index: gdcmFileHelper.cxx
>>> ===================================================================
>>> RCS file: 
>>> /cvsroot/Insight/Insight/Utilities/gdcm/src/gdcmFileHelper.cxx,v
>>> retrieving revision 1.8
>>> diff -u -3 -p -r1.8 gdcmFileHelper.cxx
>>> --- gdcmFileHelper.cxx  20 Feb 2006 21:35:24 -0000      1.8
>>> +++ gdcmFileHelper.cxx  21 Feb 2006 15:32:18 -0000
>>> @@ -1639,12 +1639,12 @@ void FileHelper::CheckMandatoryElements(
>>>
>>>     // Patient Orientation
>>>     // Can be computed from (0020|0037) :  Image Orientation (Patient)
>>> -   gdcm::Orientation o;
>>> -   std::string ori = o.GetOrientation ( FileInternal );
>>> -   if (ori != "\\" && ori != GDCM_UNFOUND)
>>> -      CheckMandatoryEntry(0x0020,0x0020,ori);
>>> -   else
>>> -      CheckMandatoryEntry(0x0020,0x0020,"");
>>> +   //gdcm::Orientation o;
>>> +   //std::string ori = o.GetOrientation ( FileInternal );
>>> +   //if (ori != "\\" && ori != GDCM_UNFOUND)
>>> +   //   CheckMandatoryEntry(0x0020,0x0020,ori);
>>> +   //else
>>> +   //   CheckMandatoryEntry(0x0020,0x0020,"");
>>>
>>>     // Modality : if missing we set it to 'OTher'
>>>     CheckMandatoryEntry(0x0008,0x0060,"OT");
>>> _______________________________________________
>>> Insight-developers mailing list
>>> Insight-developers at itk.org
>>> http://www.itk.org/mailman/listinfo/insight-developers
>>
>>
>>
>> _______________________________________________
>> Insight-developers mailing list
>> Insight-developers at itk.org
>> http://www.itk.org/mailman/listinfo/insight-developers
>>
>