<![CDATA[This study aims to develop and evaluate a deep learning-based classification model to detect periodontal and non-periodontal diseases from dental panoramic x-ray images. The dataset used consists of panoramic images processed through a data augmentation process to increase diversity, then divided into training, validation, and testing subsets to ensure the generalization ability of the model. A transfer learning model with ResNet50 architecture was applied to utilize the optimal feature extraction capability of the medical image data. The evaluation results show that the model can distinguish between the two classes with fairly good performance, although there are indications of class bias that require further refinement. Several steps such as dataset balancing, model fine-tuning, and additional data augmentation are recommended to improve generalization and prediction accuracy. With further validation, the model is expected to become an efficient and accurate tool to support clinical analysis in periodontal disease diagnosis. Translated with DeepL.com (free version)]]>
Copyrights © 2025
