[Insight-users] Segmentaion Runtime question

Luis Ibanez luis.ibanez at kitware.com
Sat Feb 23 20:33:26 EST 2008



Hi Jim



A) Roughly speaking, you should expect between 5 and 20 minutes.


B) Strictly specking, it depends on how many iteration you selected,
    and what RMS error you selected.  The run time is linearly
    proportional to the maximum number of iterations, unless it
    reaches the RMS error lower than your setting before exhausting
    the maximum number of iterations that you set.

    Note also that, your smoothing filter may also have a number
    of iterations. What smoothing filter are you using ?


You may find useful to connect an Observer to the filter, and
to listen to the iteration events. It will give you an idea
of how much you can expect. Also, if you print the RMS error
at every iteration, that will show you how close/far you may
be from converging.

You will find many examples on how to connect observers in
the ITK Software Guide


     http://www.itk.org/ItkSoftwareGuide.pdf


As a generic advice. The GeodesicActiveContour is not the filter
that you may want to use for getting a segmentation from scratch.

Instead, you want to reserve it as a "refinement" filter that must
be feed with a rough segmentation obtained by faster methods.


You may find useful to play with the SNAP application, to get
a feeling on the performance of LevelSet methods in you image
size, without having to write the code first.

You can download binary versions of SNAP from :

          http://www.itksnap.org/download/snap/


You can also get the source code from InsightApplications.

Note that the source code at www.itksnap.org may be more recent
that the one in InsightApplications.


   Regards,


     Luis


----------------------------
abatzis at ceid.upatras.gr wrote:
> 
> 
> 
> 
> 
>    Dear Fellas,
> 
> roughly speaking how long does it take so as to run
> GeodesicActiveContourLevel Set algorithm, including smoothing on a 256 x
> 192 x 128 mha image with a core duo 1,75Ghz cpu  and 1GB DDR memory?
> 
> Thanks in advance
> 
> Jim
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
> 


More information about the Insight-users mailing list