Using an Artificial Neural Network Model in Tooth Decay Diagnosis

Abstract

The current research aims to use the artificial neural network model in the diagnosis of tooth decay. The necessary data set was acquired from 3 groups including patients, dentists, and radiologists. Radiographic images of patients were fed as input to the DenseNet model, and the similarities and differences between the diagnosis of the proposed model and the dentist's diagnosis were evaluated. Contrast and detection of image edges were achieved by image processing techniques and filters. The key objective of this study was to develop a tooth decay detection system using DenseNet architecture as a special type of convolutional neural network (CNN). Subsequently, criteria such as precision, recall, accuracy, and F-measure of the DenseNet model were examined. According to the results, precision, recall, accuracy, and F-measure of the proposed model are equal to 83.33%, 80%, 91%, and 80.12% respectively. In general, the proposed method has a better performance than the compared methods regarding precision, transparency, detection speed, and accuracy.