Intelligent Diagnosis of Periapical Pathologies in Dental Radiographs Using an Optimized Triple-Ensemble Deep Learning Framework with Real-World Clinical Validation

Information from the abstract

Accurate detection of periapical pathologies from dental radiographs remains a challenging task due to variations in image quality, anatomical complexity, and limited availability of labeled clinical datasets. This study proposes an intelligent deep learning framework, termed Triple Ensemble Artificial Intelligence for Teeth Disease Classification System (A-TDCS), designed to improve diagnostic accuracy and robustness in dental radiography analysis. The proposed framework integrates three complementary components: optimized image augmentation using a Non-Population-Based Artificial Multiple Intelligence System (np-AMIS), adaptive image segmentation guided by Population-Based AMIS (pop-AMIS), and a heterogeneous ensemble of Convolutional Neural Networks (CNNs) combined through multiple decision-fusion strategies. A comprehensive factorial experimental design involving 450 configurations was conducted to identify optimal combinations of preprocessing, segmentation, network architecture, and fusion strategies. The resulting framework was evaluated on both benchmark datasets and real clinical datasets of periapical radiographs collected from the Postgraduate Endodontic Clinic, Faculty of Dentistry, Chiang Mai University. Experimental results demonstrate that the proposed A-TDCS-AMIS model consistently outperforms single architectures and homogeneous ensembles. On the TD-XI clinical dataset, the model achieved 98.84% accuracy, while evaluation on the independent unseen TD-XIV dataset achieved 98.21% accuracy, confirming strong generalization capability under real-world conditions. The results demonstrate that combining heterogeneous CNN architectures with adaptive optimization-based fusion significantly improves diagnostic reliability for dental radiographic analysis. The proposed framework provides a scalable and robust solution for AI-assisted clinical decision support in dental diagnostics, demonstrating its practical applicability in real clinical environments and contributing to the advancement of intelligent medical imaging systems.

Why this record is monitored

This record has an Impact Signal of 77/100 based on recency, source, collaboration, and bibliographic signals. It prioritizes monitoring and is not a judgment of research quality.

Related topics: Dental Radiography and Imaging · COVID-19 diagnosis using AI · AI in cancer detection