Article Sidebar

PDF (Tiếng Việt)
Published: 2026-03-31
Abstract Views: 0
Views PDF (Tiếng Việt): 0
DOI: https://doi.org/10.52852/tcncyh.v200i3.4776
Issue
Vol. 200 No. 3 (2026)
Section
Các bài báo
How to Cite
Hằng, Đỗ T., Đức, V. V., Hiếu, P. T., Hương, Đồng T. M., & Cường, N. Đức. (2026). Artificial intelligence-based segmentation of teeth, maxillary sinus, and mandibular canal on panoramic radiographs. Journal of Medical Research, 200(3), 63-70. https://doi.org/10.52852/tcncyh.v200i3.4776

Artificial intelligence-based segmentation of teeth, maxillary sinus, and mandibular canal on panoramic radiographs

Do Thuy Hang, Vu Viet Duc, Pham Trung Hieu, Dong Thi Mai Huong, Nguyen Duc Cuong

Main Article Content

Abstract

The study aimed to evaluate the performance of an artificial intelligence model in segmenting and identifying teeth, the maxillary sinus (MS), and the mandibular canal (MC) on panoramic radiographs. The dataset consisted of 3,817 panoramic images from adults aged 18 years and older. Anatomical structures were manually annotated and used to train a deep learning model based on UperNet with a ConvNeXtV2 Femto backbone; the data were divided into training, validation, and testing sets at a ratio of 8:1:1. Model performance was assessed using Precision, Recall, F1-score, and Intersection over Union (IoU). The results showed that, at the pixel level, the F1-score reached 91.93% for teeth, 97.61% for the maxillary sinus, and 90.79% for the mandibular canal, with a mean IoU of 87.84%. At the object level, the F1-scores for all structures exceeded 97%. The deep learning model demonstrated accurate and robust segmentation of major anatomical structures on panoramic radiographs, highlighting its potential applications in automated diagnosis, treatment planning, and dental education.

Article Details

Keywords

Panoramic radiographs, anatomical structure segmentation, artificial intelligence, deep learning

References

1. Srivastava MM, Kumar P, Pradhan L, et al. Detection of tooth caries in bitewing radiographs using deep learning. arXiv. 2017; 1711.07312.
2. Vinayahalingam S, Xi T, Bergé S, et al. Automated detection of third molars and mandibular nerve by deep learning. Sci Rep. 2019; 9(1): 1-7.
3. Ngoc VTN, Agwu AC, Son LH, et al. The combination of adaptive convolutional neural network and bag of visual words in automatic diagnosis of third molar complications on dental X-ray. Diagnostics. 2020; 10(4): 209. doi:10.3390/diagnostics10040209.
4. Viet DH, Son LH, Tuyen DN, et al. Comparing the accuracy of two machine learning models in detection and classification of periapical lesions using periapical radiographs. Oral Radiol. 2024; 40(4): 493-500. doi:10.1007/s11282-024-00759-1.
5. Mai Thị Giang Thanh, Võ Trương Như Ngọc, Ngô Văn Toàn, và cs. Xây dựng hệ dữ liệu hỗ trợ máy học để chẩn đoán sàng lọc sâu răng giai đoạn sớm. Tạp chí Nghiên cứu Y học. 2021; 137(1): 190-197. doi:10.52852/tcncyh.v137i1.40.
6. Cha JY, Yoon HI, Yeo IS, et al. Panoptic segmentation on panoramic radiographs: deep learning-based segmentation of various structures including maxillary sinus and mandibular canal. J Clin Med. 2021; 10(12): 2577.
7. Aung MTZ, Lim SH, Han J, et al. Deep learning-based automatic segmentation of the mandibular canal on panoramic radiographs: a multi-device study. Imaging Sci Dent. 2024; 54(1): 81-91. doi:10.5624/isd.20230245.
8. Bağ İ, Bilgir E, Bayrakdar İŞ, et al. An artificial intelligence study: automatic description of anatomic landmarks on panoramic radiographs in the pediatric population. BMC Oral Health. 2023; 23: 764. doi:10.1186/s12903-023-03532-8.
9. Şahan Keskin A, Eninanç İ. Segmentation of airways and soft tissues on panoramic radiographs using artificial intelligence technology. BMC Oral Health. 2025; 25: 876. doi:10.1186/s12903-025-06187-9.