[Insight-users] Radius of Gaussian?
Neuner Markus
neuner.markus at gmx.net
Mon Sep 14 08:53:33 EDT 2009
Hi,
From the documentation at
http://www.itk.org/Doxygen/html/classitk_1_1GaussianSpatialObject.html#2b02903cc49891d37937a9bd62ec08dd
"The Radius determines the bounding box, and which points are considered
to be inside the SpatialObject
<http://www.itk.org/Doxygen/html/classitk_1_1SpatialObject.html>. All
points with z-score less than the radius are in the object."
The Radius is defined through the z-score: "All points with z-score less
than the radius are in the object."
Since your maximum is 1 and your radius is 512 this does not make sense
because all z-scores all smaller than 1 and inside the spatial object.
motes motes wrote:
> How should the radius of the gaussian be understood? I have done this:
>
> const unsigned int gaussianDim = 1;
> typedef itk::GaussianSpatialObject<gaussianDim> GaussianType;
> typedef itk::Point<double, gaussianDim> GaussianPoint;
>
> const static int gaussian_radius = 512;
> const static int gaussian_max = 1;
> GaussianType::Pointer m_Gaussian = GaussianType::New();
> m_Gaussian->SetMaximum(1);
> m_Gaussian->SetRadius(512);
>
> GaussianPoint pos;
> for (int i=0; i<100; i++) {
> pos[0] = -300 + i*5;
>
> double weight = 0.0;
> m_Gaussian->ValueAt(pos, weight);
> std::cout << "pos = " << pos << std::endl;
> std::cout << "weight = " << weight << std::endl;
> }
>
> But I only get non-zero values in the range [-35, 35]:
>
> pos = [-35]
> weight = 9.87711e-267
> pos = [-30]
> weight = 3.69388e-196
> pos = [-25]
> weight = 1.91856e-136
> pos = [-20]
> weight = 1.3839e-087
> pos = [-15]
> weight = 1.38634e-049
> pos = [-10]
> weight = 1.92875e-022
> pos = [-5]
> weight = 3.72665e-006
> pos = [0]
> weight = 1
> pos = [5]
> weight = 3.72665e-006
> pos = [10]
> weight = 1.92875e-022
> pos = [15]
> weight = 1.38634e-049
> pos = [20]
> weight = 1.3839e-087
> pos = [25]
> weight = 1.91856e-136
> pos = [30]
> weight = 3.69388e-196
> pos = [35]
> weight = 9.87711e-267
> pos = [40]
>
>
> When the radius is 512 I thought that the bell curve would have
> support in the interval: [-512, 512] but it seems that the radius in
> the gaussian expresses something else.
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the ITK FAQ at: http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>
>
More information about the Insight-users
mailing list