[Dcmlib] Re: Troubles again on jpeg2000

[Mathieu Malaterre]

Since I don't have enough hours in the day...

For whoever is interesed in decompressing DICOM image encoded with 
*irreversible* transfer function, you have to unappkly this patch:

http://www.creatis.insa-lyon.fr/viewcvs/viewcvs.cgi/gdcm/src/gdcmjasper/src/libjasper/jpc/jpc_fix.h?rev=1.3&sortby=date&view=log

This patch will allow one to see properly (well almost) images like 
CT_Phillips_JPEG2K_Decompr_Problem.dcm

I cannot get enough attention on jasper ML, even people from dcmtk are 
banging there head on this problem. jasper simply does not handle 16bits 
image in the irrersible case (well at least this one).

A much better library look like kakadu. I would suggest for real serious 
people to use this library instead. Here is how:

Uncomment the #if 0 in gdcmJpeg2000.cxx so that when reading a DICOM 
image it will create a temp file containing only the jpeg2000 stream.
Then use kakadu to extract the image:

$ kdu_expand /tmp/jpeg2000.jpc /tmp/output.bmp

I know this is a pain, but I cannot garantee anything otherwise.

Mathieu
Ps: CT_Phillips_JPEG2K_Decompr_Problem.dcm contains indeed a jpeg2000 
stream compressed using kakadu 2.2 !

Mathieu Malaterre wrote:
> Jean-Pierre Roux wrote:
> 
>> Hi, Mathieu
>>
>> Remember the jpeg2000 images with 2 hypointense shapes  (an a strange 
>> rectangle on the top of the image  -but MD said it was a CR artifiact, 
>> no a decompression pb)?
>> You made some modif in JASPER, and it was OK (no more shapes, still 
>> the rectangle)
>>
>> The pb is here again (even with 'old style' TagKeys)
>> JP
> 
> 
> http://public.kitware.com/GDCM/Testing/Dashboard/20050711-0100-Nightly/Update.html 
> 
> ->
> http://www.creatis.insa-lyon.fr/viewcvs/viewcvs.cgi/gdcm/Testing/CMakeLists.txt.diff?r1=1.112&r2=1.113 
> 
> 
> Doing my best...
> Mathieu
> _______________________________________________
> Dcmlib mailing list
> Dcmlib at creatis.insa-lyon.fr
> http://www.creatis.insa-lyon.fr/mailman/listinfo/dcmlib
>