Detecting Dental Caries Using Oral Imagery Based on Deep Learning Algorithms: A Systematic Review

Oral diseases such as dental caries are distressing health problems that affect people throughout their lifetime and they require early diagnosis to avoid extreme measures and expenditure. With development in the application of Artificial Intelligence, Deep Learning (DL) has evaloved as a potential technology in dental diagnosis bringing about efficiency, precision and reproducibility in caries diagnosis. DL models such as Convolutional Neural Networks (CNNs), these techniques show outstanding performance in distinguishing carious tissue and its extent, detecting and mapping carious lesions across different imaging techniques. However, there are some challenges in this field, namely the dependency on high-quality annotated data, the susceptibility to variability of images, and problems with early and secondary caries identification. The current review aims at categorising the existing literature and comparing the DL-based methods for detecting dental caries and identifies the key research gaps. This review determines the existing methodologies and their usage in detail while discussing the challenges. This research helps researchers to understand the state-of-the-art DL models and discuss future directions for concerning issues like dataset variety, model transferability, and their implementation into clinical practices.