Panoramik Radyograflarda Diş Tiplerinin Sınıflandırılması için Derin Öğrenme Yöntemlerinin Karşılaştırılması

Authors : Berrin Çelik, Mehmet Zahid Genç, Mahmut Emin Çelik

Pages : 87-95

View : 35 | Download : 35

Publication Date : 2024-01-30

Article Type : Research Paper

Abstract :Tooth type classification is routinely performed in diagnosis, treatment, and planning. Towards digital dentistry, it is valuable to perform processes automatically instead of time-consuming conventional approaches. This work suggests a novel deep learning model to classify tooth types. This study proposes a new model, ZNet, for classifying tooth types and compares its performance with the leading deep learning models. The tooth types in panoramic images are categorized into 4 classes: incisor, canine, premolar and molar. This study investigates the performance of 7 different deep learning models, namely ResNet-50, VGG-19, EfficientNet, Densenet, Inception, Xception and the proposed ZNet. Model performances are evaluated using Accuracy, Precision, Recall and F1-score metrics. Accuracy, Precision, Recall and F1-score for the proposed model ZNet are 95.79%, 84.10%, 94.80% and 87.60% respectively. The proposed model outperformed the others. Findings showed deep learning models have been shown to be reliable tools providing accurate predictions in the classification of tooth types.
Keywords : diş, sınıflandırma, yapay zeka, derin öğrenme, panoramik radyografi