The Application of Artificial Intelligence for Tooth Segmentation in CBCT Images: A Systematic Review

Tarce, Mihai and Zhou, You and Antonelli, Alessandro and Becker, Kathrin (2024) The Application of Artificial Intelligence for Tooth Segmentation in CBCT Images: A Systematic Review. Applied Sciences, 14 (14). p. 6298. ISSN 2076-3417

[thumbnail of applsci-14-06298.pdf] Text
applsci-14-06298.pdf - Published Version

Download (1MB)

Abstract

Objective: To conduct a comprehensive and systematic review of the application of existing artificial intelligence for tooth segmentation in CBCT images. Materials and Methods: A literature search of the MEDLINE, Web of Science, and Scopus databases to find publications from inception through 21 August 2023, non-English publications excluded. The risk of bias and applicability of each article was assessed using QUADAS-2, and data on segmentation category, research model, sample size and groupings, and evaluation metrics were extracted from the articles. Results: A total of 34 articles were included. Artificial intelligence methods mainly involve deep learning-based techniques, including Convolutional Neural Networks (CNNs), Fully Convolutional Networks (FCNs), and CNN-based network structures, such as U-Net and V-Net. They utilize multi-stage strategies and combine other mechanisms and algorithms to further improve the semantic or instance segmentation performance of CBCT images, and most of the models have a Dice similarity coefficient greater than 90% and accuracy ranging from 83% to 99%. Conclusions: Artificial intelligence methods have shown excellent performance in tooth segmentation of CBCT images, but still face problems, such as the small size of training data and non-uniformity of evaluation metrics, which still need to be further improved and explored for their application and evaluation in clinical applications.

Item Type: Article
Subjects: European Repository > Multidisciplinary
Depositing User: Managing Editor
Date Deposited: 20 Jul 2024 10:08
Last Modified: 20 Jul 2024 10:08
URI: http://go7publish.com/id/eprint/4532

Actions (login required)

View Item
View Item