Biomedical Engineering Reference
In-Depth Information
Figure 1.
Isocontours for a medical image.
self-intersection dependency, which requires
O
N
2
in the worst case, where
N
is the number of mesh nodes.
Moreover, the image forces may be not strong enough to push the model
toward the object boundary. (Appendix A discusses recent proposals to address
this problem.) Even the balloon model in Eq. (1) cannot deal with such problems
because it is difficult to predict if the target is inside or outside the isosurface (see
Figure 1). This makes it harder to accurately define the normal force field.
Besides, due to image field inhomogeneities, topological defects (holes) may
corrupt the extracted surface. Also, the objects may split and/or merge during the
isosurface extraction process.
These problems can be addressed through an efficient pre-segmentation. For
instance, when reconstructing the geometry of the human cerebral cortex, Prince at
al. [19] used a fuzzy segmentation method (
Adaptive Fuzzy C-Means
) to obtain the
following elements: a segmented field that provides a fuzzy membership function
for each tissue class; the mean intensity of each class; and the inhomogeneity of
the image, modeled as a smoothly varying gain field (see [19] and the references
therein).
The result can be used to steer the isosurface extraction process as well as
the deformable model, which is initialized by the obtained isosurface. We used a
similar approach in [20].
In [21], pre-segmentation is done by using the
Image Foresting Transformation
(IFT). The isosurface extraction is then performed. The IFT is a segmentation
process based on the idea of mapping an input image into a graph, computing
the shortest-path forest in this graph, and outputting an annotated image, which is
basically an image and its associated forest. A watershed transform by markers and
other connected operators can be efficiently implemented by the image foresting
transformation, which emphasizes the capabilities of this technique [23].
