[Insight-users] ImageBoundaryCondition and ImageBoundaryFaceCalculator

[Julien Michel]

Dear ITK users,

I am not sure to understand properly the mechanism to handle the 
boundary condition, when it comes to neighborhood processing loops.

I understand that the ImageBoundaryCondition objects allows to manage 
the pixels values when the region border is also the border of the 
buffered region.

What I do not understand is why using the ImageBoundaryFaceCalculator. 
This object seems to break the initial region into five pieces: the 
center, and the four stripes on the boundary of the requested region, 
with a width of the radius used in the neighborhood.

However, these five pieces seems to be processed exactly the same way in 
the examples I have seen in the code. There is no special code for the 
center region, or for the little stripes. So why breaking the requested 
region into pieces if processing each of these regions the same way ?

Thanks a lot for your response,

Best regards,

Julien