[Insight-users] The Bone Segmentation

[Luis Ibanez]

Hi Amir,


1) Your input image: is it a CT modality ?


2) What is the resolution of the image ?
    (pixel spacing)


3) How are you initializing the LevelSet in SNAP ?
    are you doing an initial manual segmentation ?


4) It is unlikely that you will manage to use the
    same parameters for segmenting the tibia and
    the femur. This is due to the fact that the
    bones have different shapes (in particular
    different curvatures), and that the distance
    between the tibia and the fibula is a lot smaller
    than the distance between the femur and its
    confronting bones (tibia and hip).


5) In general, if you are observing leaks,
    this means that:


     a) the pixel spacing is not small enough, or
     b) the image contrast is not good enough, or
     c) your generation of edge map image needs to
        be improved, or
     d) your propagation weight is too large, or
     e) you are runnning the level set for too many
        iterations, or
     f) your curvature scaling is too low, or
     g) your advection term is too low.


A screen shot of your current edge map, along with the
list of parameters that you are currently using, and a
screen shot of the current output that you are getting,
will be very useful for us to be able to advice you on
how to proceed with your segmentation.



   Regards,


      Luis



---------------------------
amir wrote:
> Hello Dear Dr. Luis Ibanez:
> 
>  
> 
> Dear, You don't Answer my question? I want to use the ITK-Snap to 
> segment the exterior boundaries of
> 
>  
> 
> the knee bones like, femur and tibia, but each parameters like advection 
> and curvature velocities are
> 
>  
> 
> changed,finally the bubble links, then can you help me or introduce the 
> parameters that I should use to
> 
>  
> 
> prevent leackage, or someone who did this process, Please.
> 
>  
> 
> Best Regards.
> 
>  
> 
> Amir Hossein Saveh.
>