[Insight-users] Problem in reading a clean DICOM file

[Mathieu Malaterre]

Salut Pef,

	[Please keep posting to the ML]

	Yes I am 200% sure that your MetaHeader is wrong. For some reason the 
two other tools you tried your DICOM file against are simply discarding 
the MetaHeader (I really wonder how they handle deflate syntax then). 
Just for reference, here is the thread, and confirmation by a third 
party that my recommendation is indeed the correct solution:
http://groups.google.com/group/comp.protocols.dicom/browse_frm/thread/1ff5bc522437262b/023ff076152e4da6#023ff076152e4da6

	I have downloaded your file:
http://dl-3.free.fr/52616e646f6d495645df9ec226e6fbbf49b595e728476fd82b3eba0f76e2cf63/c_vf1002.dcm
	and you did the correct change, this is a valid DICOM file now. But yes 
you are right, when you load that image using GDCM it looks like there 
is an issue with the pixel scalar range. I'll try to have a look at the 
issue. How exactly are you generating those image ? Do you have the 
original, do you know the scalar range for those ?

Thanks
Mathieu
Ps: In ITK the image should not appear black, are you sure you are using 
unsigned short for the itk::Image pixel type ?


pef magic wrote:
> I have tried to change the header like you said (example : 
> <http://dl-3.free.fr/52616e646f6d495645df9ec226e6fbbf49b595e728476fd82b3eba0f76e2cf63/c_vf1002.dcm 
> <http://dl-3.free.fr/52616e646f6d495645df9ec226e6fbbf49b595e728476fd82b3eba0f76e2cf63/c_vf1002.dcm>>). 
> The files are now readble by itk but give a black image.
> With the other free softwares, the images shown (after hex editing) are 
> not good : it is the same display as if there is big/little endian 
> problem. Are you sure the dataset is in little endian ?
> 
> 2006/10/24, Mathieu Malaterre <mathieu.malaterre at kitware.com 
> <mailto:mathieu.malaterre at kitware.com>>:
> 
>     pef magic wrote:
>      > Hello,
>      >
>      > I am trying to open a DICOM file (joined with the mail), taking as
>      > example, the file "examples/IO/DicomImageReadWrite.cxx".
>      > First, I have tried to open this file with other free software
>     and they
>      > seem to know it as a clean dicom file.
>      > Nevertheless, it doesn't work with itk and stops at the  "try
>      > {reader->Update();}catch" test. The same code, with other dicoms
>     I had,
>      > works perfectly, of course.
>      > The dicom files (those which do not work) had been obtained by
>      > converting a "Visible Woman" General Electric Genesis format into the
>      > "dcm" format thanks to ezDicom.exe.
>      > Do you know what is the problem ?
> 
> 
>     I think I'll have to really support this one. No this is NOT a valid
>     DICOM file. Your meta header specify that the dataset is encoded in Big
>     Endian:
> 
>     D 0002|0010 [UI]
>     [Transfer Syntax UID] [1.2.840.10008.1.2.2 ]  ==>       [Explicit VR -
>     Big Endian]
> 
> 
>     But your dataset was encoded in little endian. If you have a lot of
>     those file, then you need to open them with an hex editor and change the
>     string
> 
>     1.2.840.10008.1.2.2 into 1.2.840.10008.1.2.1
> 
>     Then try running ITK again on your file.
> 
>     Mathieu
>     Ps: just for curiosity on how many free software, please quote them, did
>     you try your image against.
> 
>