Classification of Dental Cavities from X-ray images using Deep CNN algorithm
M. Muthu Lakshmi, P. Chitra
Abstract
Nowadays the dental cavities mostly occur due to the intake of Sugary drinks, candy, and food particles. They stay in a tooth and after some time it generates bacteria. These bacteria along with the acid and saliva cause plaque formation. The plaque is targeted by our enamel of a tooth then naturally appears in blackish holes. To overcome these issues the early diagnosis of dental illness about the plague and black holes must be developed. The analysis of dental image processing allows precise selection of early dental disease. The proposed method uses the Sobel edge detection with deep CNN to predict the cavities in the early stages. The preprocessing is done by histogram equalization, enhancement of contrast, and feature selection. The proposed method is compared by using the implementation of different segmentation techniques induced with the deep CNN. The various segmentation methods used for comparison are Otsu's threshold and Watershed. The suggested method uses the Sobel method for detecting the edges in dental images which are used to measure the gradient of the pixel intensity values in direction Gx and Gy. This method is efficient for prediction compared to other methods and it achieved an accuracy of 96.08%.