Re: [Gdcm2] GDCM cannot get rescale intercept/slope/type from nested objects with undefined length

[<k3...@fr...>]

Thanks for this detailed answer.

Could you explain me what DCS and PMS stand for?

Here is the output of gdcminfo for that image:
MediaStorage is 1.2.840.10008.5.1.4.1.1.4 [MR Image Storage]
TransferSyntax is 1.2.840.10008.1.2 [Implicit VR Little Endian: Default Transfer Syntax for DICOM]
NumberOfDimensions: 2
Dimensions: (160,160,1)
SamplesPerPixel    :1
BitsAllocated      :16
BitsStored         :12
HighBit            :11
PixelRepresentation:0
ScalarType found   :UINT16
PhotometricInterpretation: MONOCHROME2 
PlanarConfiguration: 0
TransferSyntax: 1.2.840.10008.1.2
Origin: (-73.2212,-67.9772,129.333)
Spacing: (1,1,3)
DirectionCosines: (1,0,0,0,1,0)
Rescale Intercept/Slope: (-3142,1.53455)
Orientation Label: AXIAL

I guess we should blame ITK right? :)
Thanks

Florent

----- Mail original -----
De: "Mathieu Malaterre" <mat...@gm...>
À: k3...@fr...
Cc: "gdcm-developers" <gdc...@li...>
Envoyé: Jeudi 17 Août 2017 14:18:29
Objet: Re: [Gdcm2] GDCM cannot get rescale intercept/slope/type from nested objects with undefined length

On Thu, Aug 17, 2017 at 2:10 PM, Mathieu Malaterre
<mat...@gm...> wrote:
> Salut Florent :)
>
> On Thu, Aug 17, 2017 at 11:16 AM,  <k3...@fr...> wrote:
>> Hi,
>>
>> I am trying to read a DICOM series (multi-echo series from Philips) but I need to check some DICOM tags first:
>> - 0028,1052: rescale intercept
>> - 0028,1053: rescale slope
>> - 0028,1054: rescale type
>
> So you are dealing with legacy MR Image Storage instance(s) from
> Philips Medical System.
>
>> Unfortunately GDCM is not able to find that tags through ITKs API (gdcmIO->GetMetaDataDictionary()), but when I use gdcmdump to search for DICOM tags manually I am able to find the tags but in nested objects with "undefined length":
>>
>> (2001,1068) ?? (SQ) (Sequence with undefined length)          # u/l,1 ?LinearModalityGLTrafo
>>       (fffe,e000) na (Item with undefined length)
>>         (0028,1052) ?? (DS) [-12095]                              # 6,1 Rescale Intercept
>>         (0028,1053) ?? (DS) [5.90720390720390]                    # 16,1 Rescale Slope
>>         (0028,1054) ?? (LO) [normalized]                          # 10,1 Rescale Type
>>         (2001,0010) ?? (LO) [Philips Imaging DD 001]              # 22,1 Private Creator
>>         (2001,1077) ?? (CS) [LINEARMODALITY]                      # 14,1 GL TrafoType
>>         (2001,10c1) ?? (LO) [LinearModalityGLTrafo ]              # 22,1 Nested Object Type Name
>>       (fffe,e00d)
>> (fffe,e0dd)
>
> Well since the public tag is nested within a private tags, this means
> it belong to the extended SOP Class definition. In case you have
> access to the DCS of PMS for that particular instance, you should be
> able to implement support for it.
>
>> Is there a way to fix that? Did you ever see that kind of nested objects? Is there a configuration on the Philips workstation to avoid that kind of nested objects?
>> Thanks.
>
> I am not 100% convince there is anything to 'fix', could you please
> send us the output of gdcminfo on that image. This will tell whether
> the issue is really at GDCM level or at ITK level. I did implement
> support for those Philips 'Combine MR Rescale option', since GDCM
> 2.6.x

I forgot to answer the last question: yes you can avoid this kind of
object, since by definition MR Image Storage cannot store Rescale
Intercept/Slope.

See for example:
https://groups.google.com/d/msg/comp.protocols.dicom/M4kdqcrs50Y/fJVGWNsgAwAJ

-- 
Mathieu