Image Processing of Panoramic Dental X-Ray for Identifying Proximal Caries

This study aims to facilitate the identification of proximal caries in the Panoramic Dental X-Ray image. Twenty-seven X-Ray images of proximal caries were elaborated. The images in digital form were processed using Matlab and Multiple Morphological Gradients. The process produced sharper images and clarifies the edges of the objects in the images. This makes the characteristics of the proximal caries and the caries severity can be identified precisely


Introduction
Proximal caries is the caries found on the surface of the tooth facing the adjacent tooth surface in the dental arches.The most difficult caries to detect is proximal caries [1].Proximal caries is the most difficult one to be detected visually in clinical examination [2,3].Radiograph called Dental X-Ray [4][5][6] is needed to detect this caries.
Dental X-Ray is an image of dental radiographs along with bones and soft tissues in the vicinity.X-ray image can show cavities (caries), hidden teeth and bones that can not be seen in visual inspection.Dental X-Ray images used to detect dental caries consists of: a. Bitewing X-ray, showing the back of the upper and lower teeth and how the teeth contact to each other in a single view.b.Periapical X-Ray, showing the whole picture of the tooth (covering several teeth from crown to root tip) to the surrounding jaw bone.c.Panoramic X-Ray, shows a view of the entire teeth, upper and lower jaw and the sinuses, the nose and the jaw joints.Some previous studies to identify proximal caries using radiographic images can be seen in Table 1.
It can be concluded from the previous studies (Table 1) that the identification of caries proximal in Panoramic Dental X-Ray images are less favorable than that of Bitewing Dental X-Ray.Identification of proximal caries using Dental Panoramic X-Ray lowest precision was compared with both other Dental X-Ray [16].This research is aimed to improve the quality of image processing in identifying proximal caries in Dental Panoramic X-Ray images.The process of analysis of such images is important in order to improve quantify medical imaging systems [17,18].It requires image processing that could generate sharper and clearer images or improve the quality of information contained in the image so that it can be visually interpreted better [19].The image processing are very helpful in identifying objects contained in the panoramic image, especially in detecting cavities [20].Therefore, automatic feature detection can assist doctor in processing the image in analyzing disease [21,22]

Results and Analysis
This study consists of six stages of processes implemented in MatLab program and one step of manual identification analysis.The last one is the significant contribution of this study.

Pre Processing by Cropping
On the the left and right side, as well as in the upper and lower parts of the image of Panoramic Dental X-Ray, there are areas that are not required for subsequent processing.To eliminate these areas in order to facilitate the determination of proximal caries in teeth, we need to crop the image of the teeth.Because the teeth are assembled in one group, then the cropping process required two coordinate points.The first coordinate point determines the upper left position (the position of the patient's right or [R] ight) and the second one determines lower right position of the cropped area (left position patients) [23,24].The cropping process results can be seen in Figure 4.The result of cropping process on input image in Figure 3 can be seen in Figure 5.

Morphological Gradient
Morphological Gradient is implied to the cropped image to clarify the edges of each object in the image.It is done by reducing the result of dilation process with that of morphology erosion.Morphology dilation is intended to widen the area of the edges of the object.Meanwhile, the morphology erosion is aimed to discourage or dilute the edge area of the object.There will be differences in the area of each object when they are compared with the objects on the wide dilation results.The stages of the process are illustrated in Figure 6.Processing result of Morphological Gradient process is shown in Figure 7.

Multiple Morphological Gradient (mMG)
Multiple Morphological Gradient is an iterative process of Morphological Gradient, results in an image with the highest intensity that can be obtained using the Morphological Gradient [25].mMG will sharpen the boundary edges of objects.The mMG results can be seen in Figure 8.

Identification of Proximal Caries in image of mMG result
The image resulted by the mMG has clear edges object so that proximal caries and its severity can visually be better interpreted.Characteristics of proximal caries on the produced image are as follows: (1) Boundary of the side edges of teeth does not overlap.(2) Boundary of the side edges protrude into the teeth (enamel, dentin, or pulp).Figure 9 shows the results of the identification by mMG.

Conclusion
This study was conducted to overcome the difficulties in identifying proximal caries through image processing of panoramic dental x-ray image.The image processing method used in this study is Multiple Morphological Gradients.These image processing results can be used to appropriately identify proximal caries and the level of severity.
Based on the results of this study, further research is needed to compare the results of Panoramic Dental X-Ray processed by mMG with the results of bitewing Dental X-Ray from any proximal caries Patients.

Figure 1 .
Figure 1.Stages of Research


ISSN: 1693-6930 TELKOMNIKA Vol. 15, No. 2, June 2017 : 702 -708 704 standard size of 2776 × 1330 pixel.Twenty-seven images containing proximal caries were identified in this study.All images have been examined by the dentists of Semen Padang Hospital, Indonesia to detect proximal caries and its severity.The position number of teeth can be seen in Figure 2.

Figure 2 .
Figure 2. Position Number and Layer of Teeth

Figure 3 Figure 3 .
below shows three images to be processed: The Image Acquisition of Patient with Proximal Carries (a) Patient 1, (b) Patient 2, (c) Patient 3

Figure 4 .
ISSN: 1693-6930  Image Processing of Panoramic Dental X-Ray for Identifying Proximal… (Jufriadif Na`am) The process of Image Acquisition (a) Initial coordinate point is at upper left (Upper right of the patient position /[R]ight), (b) Initial coordinate point is at lower left (Under right of the patient position/[L]eft), (c) Image Cropping Result

Table 1 .
. Some Studies on Identification of Proximal Caries in Dental X-Ray Images