CNN-based Approach for Prediction of Periodontally Teeth

Periodontal disease (PD) is characterized by gingival inflammation, connective tissue loss, or alveolar bone degradation. It's the world's sixth most common inflammatory disease. Clinical examination and radiograph analysis are two of the most popular approaches used nowadays to diagnose periodontal disease. Due to the constraints of the standard diagnostic paradigm, inaccuracy due to human error, and conflicting judgements by various examiners. As a result, a more consistent method of identifying periodontal disease is required. The main contribution of this paper is to propose a deep learning approach based on CNN-architectur to diagnose and predict periodontally teeth from X-ray images. There are several types of deep convolution networks, such as Vgg16, Alexnet, Inception, Resnet50, Resnet101 and Densenet169 are used. The experimental results, on combined datasets of different resources, shows that VGG16, VGG19, and Xception are the best three models, with an accuracy score of 95%, 88%, and 84% respectively. Additionally, in terms of F-Score, that VGG16 is the best model with an F1-Score of 89.5%.