Hi Marius,

AccessTwoImagesFixedDimensionByItk has indeed to create quite a lot
of code: since the function to call is templated over two unknown image
types, the access macro creates code for all common types (which are 8:
char, short, int, and the respective unsigned versions, as well as float
and
double) *times* that number, i.e., 8x8=64!
(see
http://www.mitk.org/documentation/mitk-0.8/group__Adaptor.html#g0f12c9f206cd8e385bfaaff8afeb73c7
and click on _accessTwoImagesAllTypesByItk and then on _accessTwoImagesByItk
and then on _accessTwoImagesByItk2).

That is the reason that we rarely use this access method. Normally it is
sufficient to use the macros for a single image. So what exactly does your
ITK filter chain do? Or do you even know the pixel type(s) that your
images will have?

For larger filter chains similar problems even occur for single images.
Therefore, we have implemented  workarounds for these cases (but
currently they work only for the single image versions). One is to
explicitly instantiate the function to call in separate .cpp files for parts
of the image types (currently there are macros for splitting w.r.t. the
image dimension). There is an example in the second half of the page
describing the tutorial step 6
http://www.mitk.org/documentation/mitk-0.8/Step6Page.html
(line starts with the words "Additionally the access function has to be
instantiated").
Another workaround is to reduce the number of common types that
should be supported. There is a cmake-macro for that purpose.
Currently, this works only for the ...FixedPixelType versions of the
access methods (afaik).

Best regards,

Ivo

Marius Erdt wrote:
> Hi all,
>
> I was using one of the MITK templates like 
> "AccessTwoImagesFixedDimensionByItk" to implement some ITK filter 
> chains, when VS2005 came up with the error message
>
> Error C1128, number of sections exceeded object file format limit
>
> increasing the number of sections in the .obj. file with the /bigobj 
> command solved the problem. However, this only works with VS2005 or later.
>
> Did some of you experienced similar problems when using ITK filter 
> chains? I'm just confused, because I was not using many ITK filters 
> (like 4 or 5) and those templates should generally not cause any 
> problems. However, I don't know what MITK does with the templates 
> internally, so maybe there is a way to circumvent this behavior.
>
> Regards,
> Marius
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> mitk-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/mitk-users
>   


-- 
----------------------------------------------------------------------
Dr. Ivo Wolf, Dipl.-Phys.               Phone:  (+49) 6221/42-2327
Deutsches Krebsforschungszentrum        Fax:    (+49) 6221/42-2345
Div. Medical and Biological Informatics (E130)
Im Neuenheimer Feld 280                 email:  [EMAIL PROTECTED]
D-69120 Heidelberg, Germany             http://mbi.dkfz-heidelberg.de
----------------------------------------------------------------------



-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
mitk-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mitk-users

Reply via email to