Systematic Review of Deep Learning Models for Dental Images

Abstract

Dental practitioners use digital radiographs in conjunction with clinical diagnosis to examine dental disorders of patients. Automation in dental imaging in an effort to lighten the workload of dentists is a rising subject of study within the ever-expanding healthcare sector. Recognition of similar patterns coupled with deep learning tasks is a major boon to the dental industry in the automatic diagnosis of dental diseases by examining complex anatomy. Rather than traditional image processing and machine learning, the development of deep learning algorithms has made autonomous feature extraction possible without the assistance of a human. The current study briefly discusses the robust deep learning architectures adopted for the automation of dental imaging tasks like object detection, segmentation, and classification tasks. Along with offering fresh viewpoints on this fascinating new development in the discipline, we also point out some shortcomings such as blurry edges, overfitting and within-label variation with the need for potential improvement in the discipline of dental radiology adopting deep learning techniques to focus on content-rich information in the dental radiographs.