372
F. Shi et al.
(a)
(b)
Surface 1
Surface 2
Surface 5
Surface 3
Surface 4
Fig. 13.5 Example of layer segmentation. a Original OCT image; b surface segmentation result
with local modification
13.2.4 Meshing
The ISO2Mesh method [14] is used for conducting tetrahedral meshing on the segmented CNV volumes and the three retinal layers. It is a 3-D surface and volumetric
mesh generator that can produce high quality 3-D tetrahedral mesh or triangular
surfaces directly from segmented, binary or grayscale medical images [15].
Meshing procedure includes the following two steps: (1) Generating triangular
iso-surfaces with the specified density. (2) Filling the sub-volumes bounded by isosurfaces with tetrahedral elements. In the second step different labels can be supplied
so that the resulting FEM mesh carries sub-domains that correspond to CNV volumes or the retinal layers. Figure 13.6 shows the result of retinal volumetric mesh
generation.
13.2.5 CNV Growth Model
The reaction-diffusion model [16], which is widely used in geology, biology and
physics, is used here to characterize the growth of CNV. The model describes the
spatial distribution and temporal development of certain substance. In [17–19], it
F. Shi et al.
(a)
(b)
Surface 1
Surface 2
Surface 5
Surface 3
Surface 4
Fig. 13.5 Example of layer segmentation. a Original OCT image; b surface segmentation result
with local modification
13.2.4 Meshing
The ISO2Mesh method [14] is used for conducting tetrahedral meshing on the segmented CNV volumes and the three retinal layers. It is a 3-D surface and volumetric
mesh generator that can produce high quality 3-D tetrahedral mesh or triangular
surfaces directly from segmented, binary or grayscale medical images [15].
Meshing procedure includes the following two steps: (1) Generating triangular
iso-surfaces with the specified density. (2) Filling the sub-volumes bounded by isosurfaces with tetrahedral elements. In the second step different labels can be supplied
so that the resulting FEM mesh carries sub-domains that correspond to CNV volumes or the retinal layers. Figure 13.6 shows the result of retinal volumetric mesh
generation.
13.2.5 CNV Growth Model
The reaction-diffusion model [16], which is widely used in geology, biology and
physics, is used here to characterize the growth of CNV. The model describes the
spatial distribution and temporal development of certain substance. In [17–19], it