I've been reading about chessboard, city block and Euclidean distances (ED) in
the context of connectivity.
For 2D images, it seems rather easy to understand the Euclidean distances shared
by pixels in the same neighborhood: 4-connected: 1; 8-connected: sqrt(2),
corresponding to diagonal distances.
The doubt I'm having is how this applies to 3D, in which images have almost
always anisotropic voxels.
Does one need to take into account anisotropy when defining connectivity using
Euclidean distances?
For 6-connected regions it probably does not matter, but does anisotropy affect
18-/26-connectivity defined simply by EDs?
-tiago
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html
Locked
