[Insight-users] GeodesicActiveContours processing time and SignedDanielssonDistanceMap

Dawood Masslawi masslawi at gmail.com
Wed Feb 16 10:52:13 EST 2011

To use the SignedDanielssonDistanceMap just pipeline the result of the distance map
filter to the level set filter instead of the fastMarching filter. If you already have done that
double check your distance map settings then try to change the InsideIsPositive() property
of the distance map filter which could be the source of the problem.
The Chan-Vese level set segmentation is already supported by the ITK, for more information
take a look at the following links,
Best regards,

--- On Wed, 2/16/11, Ella Maria Kadas <ella_kadas at yahoo.com> wrote:

From: Ella Maria Kadas <ella_kadas at yahoo.com>
Subject: RE: GeodesicActiveContours processing time and SignedDanielssonDistanceMap
To: "Dawood Masslawi" <masslawi at gmail.com>
Date: Wednesday, February 16, 2011, 10:27 AM

Hi Danwood,
   Thanks a lot. No, i was not going to use both of them, i was just saying that i would have liked to try to initialize the level set using SignedDanielsssonDistanceMap, but i didn't manage to get it work.Thanks for the links i will have to tune the parameters then to make it work for my images.I will maybe play first with the ITK Snap.Do you know if there is already the Chan-Vese Method using the Level Set Framework as filter for 3D segmentation implemented.?Thanks again,Ella

--- Dawood Masslawi <masslawi at gmail.com> schrieb am Mi, 16.2.2011:

Von: Dawood Masslawi <masslawi at gmail.com>
Betreff: RE:
 GeodesicActiveContours processing time and SignedDanielssonDistanceMap
An: ella_kadas at yahoo.com
CC: insight-users at itk.org
Datum: Mittwoch, 16. Februar, 2011 15:53 Uhr

Hi Ella,
You don't need to use the fastMarching filter if you are using the DanielssonDistanceMap
and vice versa.
If you want to lower the computation time you can decrease the number of iterations or 
assign lower values for the maximum allowed RMS.
The initialDistance parameter is the distance between the seed points and the desired
initial contour.
Also, the following posts on the mailing list can give you more information on how
control the level of details in your segmentation.
Note that the assigned Alpha and Beta values for the sigmoid filter can have a significant
effect on the results. It is best to choose the Beta close to the initial contour value and 
assign the Alpha in which an acceptable range of values around the Beta would be


I managed to start run the GeodesicActiveContoursFilter. I forgot to the line
I wanted to use the SignedDanielssonDistanceMapFilter, but i get the same problem as before putting the line above as a result.
Is there something similar for the SignedDanielssonDistanceMapFilter to set?

My second question is regarding the time used for the segmentation
I tried to use it on a MRI t2_3D, 256 slices 40 Iteration steps, but it took me more then 4 minutes. Is this a normal behaviour that i should expect on Leopard 10.6 16 GB RAM?

I still think i am missing something because i don't see a connection between the image and the segmented region. (Which leads me to my third question, what does the DistanceValue in the fastMarching means- it just doesn't make sense to me when starting with a seed point only and not with some initial contour).
I send some images maybe this will help to help me:P
i didn't put all of them, but Test4 is the one where i put my seed point.https://picasaweb.google.com/ellaMkadas/TestLevelSet?authkey=Gv1sRgCMrzgp3nsPfWeg#


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20110216/66a94eb4/attachment.htm>

More information about the Insight-users mailing list