Document Type

Article

Publication Date

Winter 1-1-2026

Abstract

Aim: The aim of this study is to investigate the effectiveness of the Attention U-Net architecture for teeth segmentation in panoramic dental images. By exploring the potential of Attention U-Net, the study aims to contribute to the development of more precise and automated dental imaging analysis, which could enhance diagnosis, treatment planning, and computer-aided dental systems. Subjects and methods: The TUFTS University dataset consisting of 1000 de-identified panoramic radiographs was used for training and testing an Attention U-Net architecture with 10-fold cross-validation. These scores surpass those of all other networks implemented on the same dataset. using a newly available benchmark dataset called the Tufts Dental X-ray Dataset. Results: The results demonstrated the superior performance of this network in accurately segmenting the teeth. The trained model achieved an average Dice coefficient of 95.01%, intersection over union of 90.6%, and pixel accuracy of 98.82%. Conclusion: This paper presents a proof of concept for the Attention U-Net architecture applied for teeth segmentation in panoramic radiographs. These findings contribute to ongoing efforts to develop automated systems to assist dental professionals in their clinical practice

Download

Share

COinS