ELNN: A Novel Approach to Identify Periodontal Disease From Panoramic Radiographs in Earlier Stages Using Enhanced Logical Neural Networks

Periodontal disease is a prevalent condition that can lead to serious dental complications if not detected and treated early. In this paper, we propose an Enhanced Logical Neural Networks (ELNN) model for identifying periodontal disease at early stages from panoramic radiographs. The ELNN model integrates AlexN et, a convolutional neural network (CNN) for feature extraction, with XGBoost for classification, creating a robust hybrid architecture. The dataset used includes 17,654 labeled images collected from hospital records over the last two years. Our preprocessing techniques involve resizing, grayscale conversion, noise reduction, and segmentation to ensure optimal data quality for feature extraction. The ELNN model outperforms traditional machine learning models, such as Logistic Regression, Support Vector Machine (SVM), and standalone deep learning models, achieving an accuracy of 97.77<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>. The model demonstrates superior precision (96.45<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>) and recall (96.89<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>) compared to other approaches, ensuring high reliability in distinguishing early-stage periodontal disease from healthy and advanced cases. The results suggest that the ELNN model has strong potential for clinical applications in dental diagnostics, especially for early detection, which is critical for effective treatment.