[Insight-users] watershed -- Repeated with different results (history dependent?)

[Miller, James V (Research)]

What versions of the Watershed code are you using?

There was a problem with the internal caching of results that I fixed 7
weeks ago.

Jim

-----Original Message-----
From: Atwood, Robert C [mailto:r.atwood at imperial.ac.uk]
Sent: Monday, April 19, 2004 1:06 PM
To: ITK Users (E-Mail)
Subject: [Insight-users] watershed -- Repeated with different results
(history dependent?)



Hi,

If I use the watershed in a loop, to get a series of segmentations ;
then choose one of the level values and run a single filter, I do not
seem to obtain the same results as I did within the loop for the same
parameters (level and threshold) (Also I get this behaviour if I do not
change the threshold at all i.e. it is related to the level setting)


I am guessing that this has something to do with the caching of the
initial watershed regions as described in the user guide p359:


"The relabeler object at the end of the
mini-pipeline uses the hierarchy and the initial segmentation to produce
an output image at any
scale below the user-specified maximum. Data objects are cached in the
mini-pipeline so that
changing watershed depths only requires a (fast) relabeling of the basic
segmentation."

In order to make efficient use this behaviour, I structured a loop so
that the "level" is progressively decreased, so the first update of the
watershed filter uses the highest value of the "level" desired.
However,runnign the routine again with a single value chosen from among
those tested does not result in the same output. I also tried updating
the watershed once with the highest 'level' value, then changing the
value to the target value.
Any hints?

Thanks
Robert 
_______________________________________________
Insight-users mailing list
Insight-users at itk.org
http://www.itk.org/mailman/listinfo/insight-users