[Insight-users] Re: Fwd: Re: anisotropic filter - segmentation fault

[Luis Ibanez]

Hi Ghassan,


The reason for the segmentation fault is that you
are actually instantiating an

     "AnisotropicDiffusionImageFilter"

http://www.itk.org/Doxygen/html/classitk_1_1AnisotropicDiffusionImageFilter.html



This is the base class for a family of filters
and it is not intended to be instantiated.


You may want to instantiate any of its derived classed:

CurvatureAnisotropicDiffusionImageFilter
http://www.itk.org/Doxygen/html/classitk_1_1CurvatureAnisotropicDiffusionImageFilter.html

GradientAnisotropicDiffusionImageFilter
http://www.itk.org/Doxygen/html/classitk_1_1GradientAnisotropicDiffusionImageFilter.html

VectorCurvatureAnisotropicDiffusionImageFilter
http://www.itk.org/Doxygen/html/classitk_1_1VectorCurvatureAnisotropicDiffusionImageFilter.html

VectorGradientAnisotropicDiffusionImageFilter
http://www.itk.org/Doxygen/html/classitk_1_1VectorGradientAnisotropicDiffusionImageFilter.html


I just tried your code with

itkCurvatureAnisotropicDiffusionImageFilter

and works ok.


Please note that a random image is probably not the best
input for this type of filter since this is an edge preserving
smoothing. The random image is 'edges-everywhere' so the output
will look as random as the input.

You may be in a more interesting case using the BrainWeb
images from the ftp server.

Probably the best advice about iteration, conductance an
timestep is what Josh recommended on a previous email:

http://www.itk.org/pipermail/insight-users/2002-March/000169.html


 > For parameters, start with about 10-15 iterations of the filters using
 > Conductance parameters of 1.0 and TimeStep of 0.125.  You can lower
 > the time step value, but values much higher than 0.125 will not
 > produce stable solution



Please let us know if you have further questions.


Thanks



    Luis



======================================
Ghassan Hamarneh wrote:

> Thanks Luis,
> 
> I will try what you suggested.
> 
> Strange that you can't access the website! I attached the files for you.
> 
> Yes I am getting a segmentation fault:
> 
> [itk_anldf]$ ./itk_anldf ../../imgs/img1.mhd new.img 2 2
> Segmentation fault (core dumped)
> 
> I tried different iterations and conductance parameters. What range for
> the latter do you recommend?
> 
> 
> Thank you,
> /Ghassan
> 
> 
> On Fri, 6 Dec 2002, Ghassan Hamarneh wrote:
> 
>