Histórico e evolução da inteligência artificial em análise de imagens radiográficas

Abstract

This work presents a literature review on the application of Artificial Intelligence in dental radiograph analysis, with emphasis on panoramic radiographs. First, core concepts of AI, Machine Learning, and Deep Learning are discussed, as well as their relationship with computer vision tasks. Next, the historical evolution of approaches is summarized, ranging from traditional digital image processing methods to modern deep neural network architectures. The reviewed evidence indicates that deep learning methods tend to be more robust to contrast variations, noise, and anatomical superposition, with instance segmentation standing out as a key component for downstream applications such as tooth segmentation and numbering. It is concluded that, despite technical advances, broader clinical adoption still depends on more diverse annotated datasets, external validation, and prospective evaluation.