|
KMID : 1146620230120020043
|
|
Journal of Korean Academy of Advanced General Dentistry
2023 Volume.12 No. 2 p.43 ~ p.74
|
|
|
National Research Foundation of Korea Grant
|
|
Yang Su-Jin
|
|
|
Abstract
|
|
|
The aim of this literature review is to investigate current application and diagnostic performance of AI in the dental field, address their limitations, and suggest possible future applications to AI in dentistry. Studies implementing deep learning in the dental field were searched, identified, and extracted from the electronic databases (PubMed, Cochrane Library, Scopus). Full-text articles describing the application of deep learning for the detection, classification, diagnosis, or clinical outcomes of dental problems as well as the study methods and deep learning architecture were included. The initial electronic search identified 1226 titles, and 115 studies were eventually included in the review.
According to the evaluation criteria the studies all involved deep learning methods in dentistry (published 2016-2021), and the studies were divided into their scope of each subfield in dentistry; namely, anatomy (n=26), orthodontics (n=12), oral and maxillofacial surgery (n=29), endodontics and conservative dentistry (n=17), periodontology (n=5), implant dentistry (n=7), prosthodontics (n=3), forensic dentistry and identification (n=8), and etc. (n=8). There was a high risk of bias and applicability concerns were detected for most studies, mainly due to data selection and reference test conduct. Application of deep learning proposed in the studies exhibited wide clinical applications in the dental field. However, the evaluation criteria for the efficacy of deep learning have still not been clarified, and further verification of the reliability and applicability of the AI models is essential to implement these models to clinical practice.
|
|
|
KEYWORD
|
|
|
Deep learning, Machine learning, Computer vision, Dentistry, Reveiw
|
|
|
FullTexts / Linksout information
|
|
|
|
|
Listed journal information
|
|
|
|
|