[Insight-users] Who have experience using itkShapePriorSegmentationLevelSetImageFilter

[Zachary Pincus]

I would suggest closely studying the example code for using the  
ShapePriorSegmentationLevelSetImageFilter. You can find it in  
Insight/Examples/Segmentation/ 
GeodesicActiveContourShapePriorLevelSetImageFilter.cxx.

This code is documented in the new Insight 2.0 manual (in beta  
currently):
http://prdownloads.sourceforge.net/itk/ItkSoftwareGuide-2.0.0.pdf? 
download
in the Level Set Segmentation section.

Finally, it is pretty important to understand precisely how the shape  
model level set concept works. I strongly suggest reading Leventon's  
original paper on the matter:
http://splweb.bwh.harvard.edu:8000/pages/papers/leventon/cvpr00/ 
cvpr00.pdf

Now to your questions:

> 1 What model should be TInputImage?

TInputImage is the type of the initial level set that you provide to  
the filter. It would be worth familiarizing yourself with the general  
way in which ITK level set filters work, by reading the appropriate  
section of the manual.

> 2 What other parameters need to defined?

Quite a few. See the example code and the manual. You'll probably need  
to use the PCAShapeModelEstimator class to estimate a shape model from  
a set of distance maps. You'll need to use the signed distance map  
filter to turn a set of binary shape masks into a these distance maps.  
There's a lot of other stuff there too.

> 3 Is it nessary to connecting other filters as
> inputing like ThresholdLevelSetSegmentationFilter?

It depends on what you want to do! You will need to provide some  
initial level set to the filter; where it comes from is your choice.

Good luck, and I'm happy to answer any specific questions you have once  
you've gotten a chance to absorb Leventon's paper and the ITK manual  
chapter on level sets!

Zach


>
> I really appreciate your help!
>
> Thanks,
> Guowei He
>
>
>
>
> --- Zachary Pincus <zpincus at stanford.edu> wrote:
>> I've been using (and modifying) this filter
>> extensively.
>>
>> Zach Pincus
>>
>> Department of Biochemistry and Program in Biomedical
>> Informatics
>> Stanford University School of Medicine
>>
>>
>> On May 23, 2005, at 12:23 PM, he guowei wrote:
>>
>>> Hello,
>>> I tried to implemtent Leventon's "statistical
>> shape
>>> influence in Geodesic Active Contour". An I found
>> the
>>> tkShapePriorSegmentationLevelSetImageFilter in ITK
>>> have implemented this paper. Who have used the
>> filter?
>>> Could I ask some questions about this filter?
>>>
>>> Thanks,
>>> Guowei He
>>>
>>> _______________________________________________
>>> Insight-users mailing list
>>> Insight-users at itk.org
>>> http://www.itk.org/mailman/listinfo/insight-users
>>>
>>
>>
>