Smart Periodontitis Detection: An IoT and AI-Driven Approach

Sammanfattning

The study investigates an innovative approach to early detection of periodontitis using a combination of Internet of Things (IoT) and Artificial Intelligence (AI). Periodontitis is a chronic inflammatory oral disease which affects a large percentage of all adults, the disease suffers from late and invasive diagnosis methods. To overcome these limitations, this research develops and evaluates an IoT-based diagnostic system. The system leverages a synthetic biomarker dataset generated via Variational Autoencoders (VAEs), that has been tailored specifically for salivary biomarkers. The study investigates the performance of different machine learning models such as Random Forest, XGBoost, K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Logistic Regression (LR), and Deep Neural Networks (DNN). The models are evaluated based on accuracy, precision, recall, and F1-Score to identify the most effective methods for accurate and reliable periodontitis diagnosis. The model evaluation indicates that ensemble methods such as RF and XGBoost perform well in the context of periodontitis diagnosis. Furthermore, the study compares the deployment of diagnostic processing strategies in two different IoT architectures, edge computing and cloud computing. Simulations were conducted to assess accuracy, latency, and computational efficiency. Results indicate that both edge and cloud are feasible solutions for periodontitis diagnosis. However, a hybrid edge-cloud architecture could be more beneficial depending on the specific context. This research contributes with valuable insights into leveraging AI-driven IoT systems to enhance periodontitis diagnostics, paving the way for scalable, efficient, and non-invasive healthcare monitoring solutions.