[Dcmlib] Re: CREATIS CVS: gdcm jpr

[Jean-Pierre Roux]

Mathieu Malaterre wrote:

> Creatis CVS User wrote:
>
>> CVSROOT:    /cvs/public
>> Module name:    gdcm
>> Changes by:    jpr    06/01/31 12:32:06
>>
>> Modified files:
>>     src            : gdcmDocument.cxx
>> Log message:
>> The automatic set to VR = "LO" for shadow elements that could be
>> is now commented out in order not to generate 'Shadow Groups' where some
>> Data Elements are Explicit VR and some other ones Implicit VR
>> (Stupid MatLab DICOM Reader couldn't read gdcm-written images)
>>
>> URL : 
>> http://cvs.creatis.insa-lyon.fr/viewcvs/viewcvs.cgi/gdcm/?view=query&branch=&who=jpr&date=explicit&mindate=2006-01-31%2011:31&maxdate=2006-01-31%2011:33 
>>
>
>
> Ok I am not sure I understand that. Are you saying that gdcm is 
> producing non homogenous data element ? 

*produced*

You get an image, in which Public groups are explicitVR, and Shadow 
groups are ImplicitVR.
(We have tons of thoose images)
If you take into account a 'shadow dict' (as stored in Dicts), since not 
*all* the shadow elements are described in the shadow dict, you will get 
(in gdcm::File), within a single shadow group, a mix of ExplicitVR and 
ImplicitVR DataElements.
At Write time, some shadow groups may be 'non VR homogeneous'

I commented out the shadow dictionaries in Dicts/CMakeLists.txt, that 
was enought to solve the pb.

(What I was talking about in my commit was a concomittant pb :
Since we know the elements 0001 - 0010 of a shadow group must always 
have a VR = "LO", we (I) forced it to LO, when it was ImplicitVR.
--> same pb occured.
I don't force any longer; the pb is solved, too.

JP

> Could you send me one image written by gdcm to check with dcmtk, but I 
> am pretty sure that this is forbidden by standard. You cannot mix 
> Explicit and Implicit(*).
>
> Mathieu
> Ps: The only case where this happen, is a very special case where the 
> MetaHeader is in Explicit(garantee by standart) and the rest is 
> described following the TS mentionned in group 0002.
>
> _______________________________________________
> Dcmlib mailing list
> Dcmlib at creatis.insa-lyon.fr
> http://www.creatis.insa-lyon.fr/mailman/listinfo/dcmlib
>