[Insight-users] "itkVectorFuzzyConnectednessImageFilter".

[Ying Zhuge]

Hi Long,

You understood this class pretty well!

Yes, you are right. I implemented "relative, vectorial scale, multiple 
objects, multipe seeds" FC in that class, but I didn't consider 
iterative case.  Please refer to algorithm kIRFOE in IEEE PAMI paper 
2002 vol 24(11) pp. 1485-1500 "relative fuzzy connectedness and object 
definition: theory, algorithms, and applications in image segmentation" 
for iterative case, which is for two objects only. we are working for 
iterative relative FC simutaneously for multiple objects.

As for 2, the problem is from the inverse matrix computation of ZERO 
matrix. I didn't consider that extreme case in my program, but it can be 
easily modified. You just need to set the inverse matrix as ZERO in 
functions SetHomogeneityMatrix and SetObjectsMatrix if the input is 
ZERO. Then the program will work fine.

Hope it helps!

Ying

Yu Long wrote:

>Hi Ying Zhuge,
>    I am now studying the itkVectorFuzzyConnectednessImageFilter, which I am not very clear with.
>    I have read your paper "Vectorial scale-based fuzzy connected image segmentation". 
>   And  I have write codes to compute the "HomogeneityMatrix" ,"ObjectsMean" and "ObjectsMatrix".
>   In my codes , the  HomogeneityMatrix is computed according to your paper.
>
>My two puzzles are:
>1 As I have tested, the itkVectorFuzzyConnectednessImageFilter have realized the algorithm "kVMRFOE" showed in your paper completely.  "kVMRFOE" realized most ideas such as "relative, scale, multiobject, multiseeds", EXCEPT the "Iterative" idea ,right?
>2 Sometimes, we can get  "ZERO" ObjectsMatrix when the image are too homogeneous such as the image attached. 
>itkVectorFuzzyConnectednessImageFilter seems can't work well at this condition. How can I resolve this?
>
>Regards,
>
>Yu Long
>
>  
>
>
> ------------------------------------------------------------------------
>
>------------------------------------------------------------------------
>
>_______________________________________________
>Insight-users mailing list
>Insight-users at itk.org
>http://www.itk.org/mailman/listinfo/insight-users
>  
>