DCMTK 3.6.0 win32 dcmcjpeg.exe

[hsoltau]

Hi all,

i use dcmcjpeg.exe with no problems on a 12 MB Mammo Image with Version 3.5.3 win32

with version 3.6.0 i get an error:

D:\dcmtk-3.6.0-win32-i386\dcmtk-3.6.0-win32-i386\bin>dcmcjpeg.exe mammo_orig.dcm
what.dcm
E: True lossless encoder: Cannot change representation, not enough data
F: no conversion to transfer syntax JPEG Lossless, Non-hierarchical, 1st Order P
rediction possible!

what did i miss?

thanks
Henning

[Andreas Thiel]

It seems a problem with your spezific Mammofile.
everthing works fine at our site.

Can you send me a download link with the image?

And / our a full dcmdump output?

[hsoltau]

hi andreas,

its funny, dcmcjpeg.exe in version 3.6.0 doesnt seem to work with all of my test files.

the files are all exported of our PACS Vendor Workstation Software, and with 3.5.3 i can compress them without errors but not with 3.6.0

greetings
Henning

[Michael Onken]

Hi Henning,

if you would like us to have a look, send a non-working image to dicom at offis dot de, then we'll have a look.

Thanks,
Michael

[hsoltau]

thanks in advance, mail sent

[Andreas Thiel]

Hi Henning,
It is an problem with the private group 0009 (which seems to be an image icon ) in your file.
There is a 2. PixelData element in it which can not be compressed.

As a workaround: dcmodify -ep US_orig.dcm before compressing.

But of course, we have to change our code to handle such problems.

[hsoltau]

OK,
thanks a lot...
in the meantime i use the 3.5.3 version of dcmcjpeg.exe in my script

Henning

[J. Riesmeier]

From the textual dump of "US_orig.dcm" I would suppose that the PixelData ("ff\d8\ff\e0\00\10\4a...") in the private sequence (0009,1010) is JPEG compressed, although the transfer syntax of the dataset is BigEndianExplicit. This is not allowed according to the DICOM standard!

The error message ("not enough data") is also clear: 96 (Rows) * 128 (Columns) * 3 (SamplesPerPixel) = 36,864 bytes, but there are only 4,066 bytes.

I didn't check why DCMTK 3.5.3 was acting differently in this regard...