[Insight-users] Segmentation Editor - pixel spacing

Joshua Cates cates@sci.utah.edu
Fri, 22 Nov 2002 17:58:54 -0700 (MST)


Hi Dean,

The filters used in this demo (underscore the word "demo") application do
not currently make use of pixel spacing.  To have the preprocessing
stage--diffusion and gradient magnitude--take spacing into account, you
will need to resample the input data so that it is isotropic.  (Note that
the watershed algorithm itself does not take derivatives or make distance
measurements on its input image.)

We do plan to modify some of the basic image processing filters (like
anisotropic diffusion) so that they include the option to scale
derivatives.  Scaling will be optional because there is disagreement about
whether it is always appropriate.  For example, one could argue that noise
introduced in the image domain should be removed in the image domain.

Hope this clarifies things.

Josh.
______________________________
 Josh Cates			
 School of Computer Science	
 University of Utah
 Email: cates@sci.utah.edu
 Phone: (801) 587-7697
 URL:   www.cs.utk.edu/~cates


On Fri, 22 Nov 2002, Dean Inglis wrote:

> Hi,
> 
> while experimenting wth the Segmentation Editor
> app, I noticed that there appears to be no way 
> to enter information about the spacing/dimensions
> of pixels/voxels.  Is this intended and if so, why?
> 
> Dean
> 
> quote from Insight/Examples/Image/Image4.cxx :
> " Medical images lacking 
> // spatial information should not be used for purposes like medical
> // diagnosis, image analysis, feature extraction, assisted radiation
> // therapy or image guided surgery. In other words, medical images
> // lacking spacing are not only useless but rather hazardous."
> 
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users@public.kitware.com
> http://public.kitware.com/mailman/listinfo/insight-users
>