The field of artificial intelligence (AI) in dentistry has been continuously evolving, with a persistent focus on radiographic feature identification and single-modality analysis. These models have demonstrated meaningful analytical capabilities and objectivity at the pixel level in identifying specific findings or anatomical structures, steadily expanding their potential as clinical decision support tools. This ongoing advancement in visual recognition models serves as a fundamental process in validating the practical utility and clinical reliability of AI within Oral and Maxillofacial Radiology (OMFR).

This lecture provides a structured review of imaging-based AI research in OMFR, tracing its development from early machine learning approaches to contemporary deep learning models. Key research themes — including detection, segmentation, and classification across various dental imaging modalities — are examined from the perspective of their clinical relevance and applicability.

As multimodal technologies gain increasing attention in the broader medical AI field, dentistry is seeing growing interest in integrated attempts to combine radiographic information with textual data while maintaining visual analytical precision. This lecture also introduces recent research trends in multimodal approaches, including Large Vision-Language Models (LVLMs).

Through this, attendees will gain an overview of the current trajectory of imaging-based AI research and a preliminary understanding of how multimodal technologies may gradually inform the diagnostic workflows of the next generation of OMFR.