Biomedical Engineering Reference
In-Depth Information
(a)
(b)
(c)
(d)
Figure 19.
Segmentation of bubbles using confidence rate spaces: (a) original image;
(b) confidence rate of an adaboost classifier trained with a bubble model; (c) geometric
enhancement of the confidence rate map; (d) resulting segmentation. See attached CD for
color version.
adaboost. Figure 20a shows the original image, while Figure 20b displays the
confidence rate of the classifier. In this case, we fed the deformable model with this
map without further improvement. Figure 20c displays the resulting segmentation.
7.
CONCLUSIONS
In this chapter we consider a novel geodesic snake formulation, termed
STOP
and GO
, for more efficient convergence to shapes. The formulation is based
on restricting the regularizing term to the last stages of the snake deformation
