next up previous
Next: ``Segement_Generation'' -- improve Up: Topology Previous: ``Homotopic_Thinning'' -- thin

 

``Homotopic_Thickening'' -- conditional homotopic thickening

 

purpose:
Thicken a binary-valued volume conditioned (set intersection) on another binary-valued volume [25,21]. Thickening means to add layers to objects in the volume without changing the connectivity (homotopy) of objects in the volume; it is the opposite of thinning. Conditioning the thickening on another volume restricts the extent of the thickening. The function thickens the input volume until it reaches a stable configuration (i.e., no further voxels can be added without changing the connectivity or going outside the extent of the conditioning volume). The function can be done in 3-D or 2-D.

Using it in conjunction with erosion, you can separate loosely connected objects without changing the shape of objects [21].

input:
Two different 8-bit binary-valued volumes.
output:
An 8-bit binary-valued volume.
parameters:
Default parameters menu is
connectivity = 3-D
# of passes = to_completion
input from vol # = 0
conditioning vol # = 1
copy output to vol # = 0
The option connectivity = '' toggles between ``3-D'' and ``2-D'' and specifies how the operation is done.
comments:
 
  1. To separate connected objects: (1) copy input volume 0 to slot 1; (2) erode volume 0 a few times to break regions apart (you must determine empirically how many erosions are necessary); (3) do a conditional homotopic thickening on volume 0, conditioned on volume 1. The result will have the objects separated, without their shapes changing.
  2. The operation ``object separation'' in the ``integrated'' category performs the object separation function described above.
  3. This is a moderately slow operation.



 

Philip Americus
The Multidimensional Image Processing Lab
Fri Aug 30 10:26:42 EDT 1996