Graphics Reference
In-Depth Information
(a)
Dividing line between maxillary tooth and mandibular tooth
(b)
Tooth area after upper and lower
boundaries are affirmed.
Fig. 4.
Dividing Line between Maxillary and Mandibular Teeth and Upper and Lower Boundaries
After the horizontal dividing line between the maxillary tooth and mandibular tooth is
defined, the upper and lower boundaries of the tooth area can be also defined by one
or two constants. These two empirical constants are respectively the deviation dis-
tance of the maxillary tooth and mandibular tooth from the middle dividing line.
Many experimental tests have proved that a better result can be obtained when the
upper and lower deviation distance are respectively set as 450 and 350, as shown in
Figure 4 (b).
2.3
To Define Left and Right Boundaries of Tooth Position through Edge
Detection
Edge detection is usually the first stage of the image processing and also one of the
classical research topics in the machine vision field. Edge detection can significantly
reduce the data volume of the image and reserve the structural information at the
same time, so it can ease the calculation burden of the system substantially. The cor-
rectness and reliability of the detection result will have a direct influence on the un-
derstanding of the objective world from the perspective of machine vision system.
Traditional edge detection algorithm is mainly based on the gradient which is es-
sentially the first order difference of the image. The pixels in the marginal area of the
image changes more sharply so they have bigger gradients. Thus the gradient-based
edge detection algorithm can achieve better effects when the interference of noise is
small. Traditional edge detection is easy to understand and its calculated amount is
small. As a result, it has been widely applied and developed in the early stage of im-
age processing.
The calculation of the gradient of an image
f(x, y)
can be shown in Equation 9.
T
∂∂
f
f
T
xy
∇
fxy
(, ) [ , ]
=
GG
=
(9)
∂∂
xy
The values of amplitude and phase can be obtained respectively after the gradient
function of the image is obtained through the above calculation, as shown in Equa-
tions 10 and 11.
1/ 2
2
2
∂
f
∂
f
(10)
∇=
f
+
∂
x
∂
y
Search WWH ::
Custom Search