Biomedical Engineering Reference
In-Depth Information
Figure 4.
The image segmentation result of a Follicular Center Cell Lymphoma (FCC)
applying robust color GVF deformable model. The left panel shows the two initial positions
for the segmentation of both nuclei and cytoplasm, the center panel shows one snapshot of
the evolving contours, and the right panel shows the final segmentation results after 125
iterations. See attached CD for color version.
The numerical solution to (30) and (31) can be solved by treating
u
and
v
as
functions of time and solving
∇
2
u
(
x, y, t
)
−
(
u
(
x, y, t
)
−
u
t
(
x, y, t
)=
µ
f
x
(
x, y
))(
f
x
(
x, y
)+
f
y
(
x, y
))
,
(32)
∇
2
v
(
x, y, t
)
−
(
v
(
x, y, t
)
−
v
t
(
x, y, t
)=
µ
f
y
(
x, y
))(
f
x
(
x, y
)+
f
y
(
x, y
))
.
(33)
In order to set up an iterative solution, let the indices
i
,
j
and
n
correspond to
x, y
,
and
t
will be derived from the iterative algorithm for (32) and (33).
Figure 4 shows the performance of applying double robust color GVF de-
formable models on segmenting both the nuclei and the cytoplasm of a Follicu-
lar Center Cell Lymphoma (FCC). Throughout our experiments, the robust color
GVF deformable model provided accurate segmentation results while maintaining
smooth boundaries.
4.
GEODESIC OR LEVEL SET-BASED DEFORMABLE MODELS
As previously mentioned, parametric deformable models offer the advantages
of fast converge speed and simple implementation. However, they are not able to
address topological changes in the image, and they therefore cannot segment mul-
tiple objects simultaneously since the deformable models themselves cannot split
or merge automatically. Furthermore, although the GVF deformable model en-
larges the capture region significantly, it still requires the initial (starting) positions
to cover at least the center of the object of interest. Otherwise, it may fail to find
the whole object if the initial positions do not include the center [29]. Geodesic
or level set-based deformable models offer a different approach than parametric
