[Insight-users] Segmentation Fault in GeodesicActiveContour

[Luis Ibanez]

Hi Geremy,

We may need more details from your experiments
in order to figure out what is wrong with your
installation.

The first suspect is that you may be feeding
inappropriate input to this application.

Typical offenses are:


1) Read 2D images.
    This application expect 3D images as input


2) Read 3D images with less than 4 slices
    The Gaussian filter requires a minimum
    of four pixels on each dimension in order
    to compute the smoothing.


3) Incorrect pixel spacing (the more likely).
    If you feed this application with images
    that have zero value for pixel spacing,
    divisions by zero may occur.  Please
    make sure that your image have a valid
    pixel spacing.


Could you give us some details on the
characteristics of your images ?


such as:

    - Image modality
    - File format
    - Image size
    - Pixel Spacing
    - Image origin


Please also, experiment using the images that
we provide on the ftp site:

ftp://public.kitware.com/pub/itk/Data/BrainWeb/

That will help us to reproduce the problem.



Thanks


   Luis


---------------------
Geremy Heitz wrote:

> Hello,
>  
> I'm running the GeodesicActiveContour demo application, and I continue 
> to get a segmentation fault when I attempt to run the Gradient Magnitude 
> operation.  I'm also getting a seg fault when (after the fast marching 
> op works) I do the Geodesic Active Contour op.  I also get segmentation 
> faults in many of the other demo applications.  I suspect there's 
> something wrong with my setup...
> 
> I'm working on a Sun Workstation running ITK version 1.4.0, and using 
> FLTK version 1.1.4.
>  
> I know this is a very generic problem, but the seg fault isn't really 
> giving me any information to go on.
> 
> Thanks,
> Geremy