Biomedical Engineering Reference
In-Depth Information
map. The theory behind RAGS is standalone and hence the region force can be
generated starting from any reasonable segmentation technique. We also showed
its simple extension to color gradients. We demonstrated the performance of
RAGS, against the standard geometric snake and the geometric GGVF snake, on
weak edges and noisy images as well as on a number of other examples.
The experimental results have shown that the region-aided snake is much
more robust toward weak edges. Also, it has better convergence quality com-
pared with both the standard geometric snake and the geometric GGVF snake.
The weak-edge leakage problem is usually caused by inconclusive edge values
at the boundaries, which makes it difficult for gradient-based techniques to de-
fine a good edge. The gradual changes do not provide sufficient minima for the
stopping function to prevent the level set accumulating in that area. The diffused
region map gives the snake an extra underlying force at the boundaries. It also
makes the snake more tolerable to noise as shown by the harmonic shape recov-
ery experiment and many of the real images. The noise in the image introduces
local minima in the stopping function preventing the standard geometric snake
to converge to the true boundary. However, for RAGS the diffused region forces
give a better global idea of the object boundary in the noise clutter and help the
snake step closer and converge to the global minima.
10.11 Further Reading
Deformable contour models are commonly used in image processing and com-
puter vision, for example for shape description [21], object localization [22], and
visual tracking [23].
A good starting point to learn about
parametric
active contours is [24]. These
snakes have undergone significant improvements since their conception, for
example see the GVF snake in [7,9]. Region-based parametric snake frameworks
have also been reported in [25-27]
The geometric model of active contours was simultaneously proposed by
Caselles
et al.
[1] and Malladi
et al.
[2]. Geometric snakes are based on the
theory of curve evolution in time according to intrinsic geometric measures of
the image. They are numerically implemented via level sets, the theory of which
can be sought in [15, 16].
