In dentistry, there is a growing interest in convolutional neural networks and machine learning models for diagnostic applications. A precise diagnosis of periodontal disease usually requires a thorough clinical and radiologic assessment. Recent research, however, has shown how deep learning may be used to construct computer-aided detection and diagnosis algorithms for dental issues using a variety of radiography sources, thereby increasing diagnostic accuracy and reliability. Convolutional neural network models may be able to assist in classifying periodontal bone loss in this case, so we tried to develop a CNN model to perform better in this aspect. We took a dataset of 100 panoramic radiograph images and augmented them. After augmentation, our final dataset consists of 38,400 images. We used the ResNet50 and MobileNetV2 models to detect periodontal bone loss and achieved 85% accuracy for the MobileNetV2 model. To make the model more accurate, we used data augmentation techniques, added extra layers, and updated the models. These techniques increased our model’s accuracy to a praiseworthy level.