[Insight-users] Isolated Connected Filter

[Luis Ibanez]

Hi Rama,

The Isolated Connected image filter searches
for a value that will separate two regions,
one region containing one of your seed points
and the other region containing the other seed
point.

The search for this value is done using a binary
search (bipartition) algorithm. The algorithm
stops when the variation of the threshold is under
the tolerance.

The reason for having this tolerance is that the
binary search may find a value that separates the
two regions but it is too far from the critical
value where the regions may have been fused.


   Regards,


     Luis

------------------------------------------
CSPL wrote:
> Hello Luis,
>  
>     I am trying to use the "itkIsolatedConnectedImageFilter" which takes 
> a parameter with the name "Tolerance". I could not get the meaning and 
> use of it from the available documentation. Can any one please provide 
> me with more information regarding this parameter?
>  
>     I have one more problem. I have resampled a PET binary mask image 
> (128x128x35) with spacing values as 4.296875/4.296875/4.25 into a CT 
> image (512x512x35) with spacing values 0.9765625/0.9765625/4.25. Now I 
> want to overlay the resampled image on to the CT image. But the mask 
> appears to have shifted. How do I overlay them on to each other?
>  
>     I was not successful in implementing the solution you suggested to 
> me the last time I have asked you for help on this issue. Please 
> indicate if I should use any other filters.
>  
>     Since after resampling both the CT image and the mask image have the 
> same properties, I thought I could map the images using pixels (not 
> physical length/space). I am unable to get any other idea to do this.
>  
>   Regards,
>     Rama Krishna. <mailto:nathkailas at hotmail.com>