[Insight-users] WaterShedSegmentation Problem

Joshua Cates cates@sci.utah.edu
Thu, 20 Feb 2003 10:58:56 -0700 (MST)


Hello Neha,

Your conductance term and diffusion iterations seem like reasonable values 
for many types of data.  The lower threshold, however, is intended to be a 
value between 0.0 and 1.0, see the Software Guide for more information 
(http://www.itk.org/ItkSoftwareGuide.pdf).  This value reflects a 
percentage of the maximum height in the input image.

Another tip is to keep the Level parameter quite low when first 
experimenting with segmentation of an image.   For many data sets we 
have experimented with, only the first 10 percent or so (Level = 0.10) of 
the hierarchy is interesting.  Keeping the level parameter low will 
decrease computation times, often significantly.


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


On Tue, 18 Feb 2003, Neha D wrote:

> 
> Hi,
> 
> I am trying to use WaterShed Segmentation Alogorithm on .raw image.
> 
> When I run that example with parameters as 
> 
> cunductance term = 0.5
> 
> diffustion iterations = 15
> 
> lower threshold = 128
> 
> the program starts running but does not end. 
> 
> Can you pls. provide me the correct parameter values ( pls. let me know how I can calculate threshold ) 
> 
> thanks,
> 
> neha
> 
> 
> 
> ---------------------------------
> Do you Yahoo!?
> Yahoo! Shopping - Send Flowers for Valentine's Day