[Insight-users] input for levelset segmentation algorithms

[Kai Li]

Hi Luis and Ivan,
  Thank you guys for your replies. I just also looked into the code of the
SparseFieldLevelSetImageFilter class. I think both of you are right.
The comment in the ::CopyInputToOutput() function indicates that the
function will figure out the zero crossings of the input image, following
a shift scaling. So I guess I just need to make the input image good for
the zero crossing processing. I think I can just pass an input with the pixels
 inside the object to be -1 and those outside the object to be 1.
Please point it out if there is anyting that is not exactly as what I
thought.

Thanks again,
Kai

On Fri, 23 Jul 2004, Luis Ibanez wrote:

>
> HI Kai,
>
> A simper way to do it is to smooth the binary image
> using a Gaussian smoothing filter, and then rescale
> the intensities to be between -0.5 and +0.5.
>
> You can do this with the filters:
>
> SmoothingRecursiveGaussianImageFilter
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1SmoothingRecursiveGaussianImageFilter.html
>
> and
>
> RescaleIntensityImageFilter
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1RescaleIntensityImageFilter.html
>
>
> Note that the level set does not have to be a distance
> funcion from your contour. Smoothing with a large Gaussian
> of large sigma may be good enough.
>
> Note that Sigma in the filter above is expressed in physical
> units (e.g. millimeteres) and not in pixels.
>
> You will find descriptions of both filters in the SoftwareGuide
>
>
>        http://www.itk.org/ItkSoftwareGuide.pdf
>
>
>
>
> Regards,
>
>
>    Luis
>
>
>
> ------------------
> Kai Li wrote:
> > After browsing the ITK classes, it seems that a possible method could be:
> > 1) first extract the contour of the binary image,
> > 2) feed the contour as input to the DanielssonDistanceMapImageFilter
> > 3) Flip the sign of the distances of the pixels within the contour,
> >    ending up with the input initial level set to the level set algorithms.
> > Not sure if there is a more direct or better way to archieve this...
> >
> > Kai
> >
> > On Thu, 22 Jul 2004, Kai Li wrote:
> >
> >
> >>Hi,
> >>  I'm considering to use one of the Levelset segmentation algorithms, say
> >>ShapeDetectionLevelSetImageFilter. As the documentation of ITK says,  one
> >>of the inputs for algorithm is the initial level set: "The initial level
> >>set is a real image which contains the initial contour/surface as the zero
> >>level set. For example, a signed distance function from the initial
> >>contour/surface is typically used..." My question is how to convert a
> >>binary image, which is generated by a preprocessing method, to the
> >>such initial level set in ITK.
> >>
> >>Any help is highly appreciated!
> >>
> >>Kai
> >>_______________________________________________
> >>Insight-users mailing list
> >>Insight-users at itk.org
> >>http://www.itk.org/mailman/listinfo/insight-users
> >>
> >
> > _______________________________________________
> > Insight-users mailing list
> > Insight-users at itk.org
> > http://www.itk.org/mailman/listinfo/insight-users
> >
>
>
>