Research on Teeth Positioning Based on Wavelet Transform and Edge Detection - Advances in Image and Graphics Technologies - page 287

2.2 Determine the Upper and Lower Boundaries of Area of Tooth by Using

Wavelet Transform

In the X-ray images, as shown in figure 1(a) , the gray value of tooth area are general-
ly high, which makes it separate from the background. And this characteristic can
provide important basis of isolating tooth area accurately.

(a) X-ray facial image (b) Lateral projection

Fig. 1. X-ray image and its horizontal projection of the face

Another essential feature of the X-ray image is that tooth are all basically in the ho-
rizontal direction, so if we do the projection in horizontal direction, the value will be
higher on the coordinates of corresponding to the tooth part due to the feature of high
degree of tooth area. And the value of other parts are relatively lower, we can call it
highland effect, as shown in figure 1(b).This provides possibilities for testing the
tooth region.
From figure 1 we can find that there are three higher grayscale areas in the
horizontal direction of the image. And the three regions can be corresponding to
nostrils, maxillary tooth area and mandibular tooth area. In order to make sure that
this characteristic has typical significance for all of the images, we make an analysis
for 50 facial X-ray images. And we get the horizontal distribution trend diagram, as
shown in figure 2 (b) (to facilitate observation, the figure is turned by 90 degree in
horizontal direction).

(a) horizontal projection (b) 50 horizontal projection

Fig. 2. X-ray images of facial horizontal projection and distribution trends