[Insight-users] Computing the difference in statistics of two regions

[Navneeth Subramanian]

David,
What you are looking for is a statistical distance metric:
http://en.wikipedia.org/wiki/Statistical_distance

If you expect your observations to be gaussian distributed, you should
try the Kullback–Leibler divergence.
http://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence

I believe this is already available in ITK (though i must admit that i
have'nt used it myself).

Hope this helps,
navneeth

On Sun, Feb 26, 2012 at 7:03 PM, David Doria <daviddoria at gmail.com> wrote:
> Using itkLabelStatisticsImageFilter, one can compute many different
> statistics in a region. Now, is there some accepted way of comparing
> the statistics of two regions to determine a "similarity"? It doesn't
> make sense to just concatenate all of these statistics for each patch
> into a vector and find the Euclidean distance between the two vectors
> (corresponding to the two patches), because the magnitudes of each
> statistic are very different. Has anyone ever seen a good way to do
> such a comparison?
>
> Thanks,
>
> David
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.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