Deep Learning-Based Classification of Dental Disease Using X-Rays

Abstract

Dental radiography is crucial for diagnosis, treatment, and quality assessment in dentistry. To enhance clinical quality, digitalized dental X-ray image analysis systems have been developed. In this study, we preprocess a dataset of dental X-ray images and evaluate treatment quality using these images. Our aim is to propose an automated clinical quality evaluation tool to aid dentists in making decisions. We employ deep learning, a form of artificial intelligence, to detect diseases in X-ray images. The dataset consists of 126 images, labeled as Normal or Affected by dental experts. Data augmentation is applied to increase the dataset size for effective training of deep learning models. A Convolutional Neural Network (CNN) architecture is constructed, comprising convolutional, max-pooling, flatten, dense, and output layers, to classify the images as Normal or Affected. The CNN model is trained on the augmented dataset to automate clinical quality evaluation. The model's performance is evaluated based on metrics such as accuracy, loss, precision, recall, and F1-score. Our method achieves an accuracy of 97.87% and an F1-score of 60%, demonstrating comparable performance to expert dentists and radiologists.