Image Processing Reference
In-Depth Information
(
a
)
(
b
)
(
c
)
Fig. 5.8.
Examples of thinning: (
b
) surface thinning by Algorithm 5.4; (
c
)axis
thinning by algorithm; both applied to the same CT image of hand.
(
a
)
(
b
)
Fig. 5.9.
Example of axis thinning. Algorithm 5.6 was applied to a CT image of
bronchus and bronchus branches extracted from a 3D chest CT image.
we design a template based upon some heuristics, we first consider three
horizontal arrangements of
3
3
voxels and then take into account the
relationship between patterns on the upper and the lower planes. This is
the typical way to obtain a 3D template for thinning.
In any algorithm, the deletability test, while keeping the topology
of a figure unchanged, is the most essential part and the most time-
consuming part of the procedure. In the algorithms presented in this
section, the deletability test was completed directly. However, various
other approaches and variations have been reported. For example, there
is a method to detect the deletability by considering arrangements of 1-
voxels on three orthogonal planes including the current voxel (called check
planes) [Tsao81, Tsao82a, Tsao82b] (Fig. 5.11).
(iv)
Connectivity
: Only the deletability test presented in the previous chapter
claims that it is applicable to all four connectivity types (6-, 18-, 18
-, and
26-connectivity). Most other methods only treat two types of connectiv-
ity, 6- and 26-connectivity. Some methods only treat the single type of
connectivity (6- or 26-connectivity).
(v)
Mode of iteration
: Thinning algorithms are classified into two major types,
the sequential type and the parallel type, according to the execution of
×
Search WWH ::
Custom Search