Hybrid Artificial Intelligence Models for Robust Dental Signal and Image Processing: Applications in Diagnosis, Treatment Planning, and Clinical Decision Support (A Narrative Review)

Abstract

Hybrid Artificial Intelligence (AI) models are rapidly transforming modern dentistry by improving the accuracy and reliability of automated analysis in dental imaging and clinical decision support. Conventional deep learning systems such as convolutional neural networks (CNNs) and transformer-based architectures have demonstrated promising results for detection and segmentation tasks; however, their performance may degrade under real-world clinical conditions due to noise, motion artifacts, metallic scatter, low-dose acquisition, limited labeled datasets, and variations in scanners and patient anatomy. Hybrid AI frameworks address these challenges by integrating classical signal and image processing techniques with deep learning-based feature learning, thereby improving robustness, generalization, and interpretability. This review summarizes recent advancements in hybrid AI methodologies applied to dental signal and image processing, including frequency-domain enhancement, wavelet-based denoising, model-based reconstruction, and deep unfolding strategies combined with CNNs, U-Net variants, and attention mechanisms. Major clinical applications discussed include caries detection, periodontal bone loss assessment, periapical lesion identification, tooth and root segmentation, cephalometric landmark detection, orthodontic treatment planning, implant planning using CBCT, and automated pathology screening. The review further highlights evaluation metrics and validation strategies commonly reported in dental AI research, such as sensitivity, specificity, F1-score, Dice coefficient, and area under the ROC curve (AUC), along with challenges related to dataset bias, explainability, and regulatory readiness. Overall, hybrid AI models offer a technically grounded pathway to enhance diagnostic reliability and workflow efficiency, supporting the development of trustworthy intelligent dental systems for real-world clinical adoption.