Automated Teeth Segmentation in Three-Dimensional Dental Images Using A Hybrid GoogLeNet-Residual Connection Architecture with Attention Mechanism
International Journal of Medical Toxicology and Forensic Medicine,
Vol. 16 (2026),
1 Dey 2026
,
Page 1-8
https://doi.org/10.22037/ijmtfm.v16.53264
Abstract
Background: Automatic tooth segmentation in three-dimensional (3D) dental images is crucial for computer-aided diagnosis, treatment planning, and dental research. However, due to the complexity of dental structures, morphological differences, variations in image quality, background noise, and other factors, segmentation remains challenging. Current deep learning methods may also struggle to preserve fine-grained structural details and identify informative regions in intricate 3D dental images.
Methods: In this study, we propose a hybrid deep learning model that combines an improved GoogLeNet, a Residual Network (ResNet), and an attention mechanism for automatic tooth segmentation in 3D dental images. First, we apply several preprocessing steps to enhance image quality and normalize the data. Then, we use an enhanced GoogLeNet architecture with additional residual connections to capture hierarchical, fine-grained features and mitigate vanishing gradients. Moreover, we design an attention mechanism to adaptively select informative dental regions while suppressing irrelevant background information, making the segmentation results more robust and informative. We test the proposed framework on a large dataset of 3D dental images from patients with various dental conditions.
Results: The proposed hybrid framework outperforms existing approaches in segmentation performance and generalization, achieving 99% Accuracy. The incorporation of residual connections helps extract detailed dental features, while the attention mechanism aids in identifying relevant dental regions and improves the interpretability of the segmentation results.
Conclusion: The proposed framework, GoogLeNet-RSNet-Attention, is an effective model for automatically segmenting teeth in 3D Dental images. The method can be used in computer-aided dental diagnosis, treatment planning, and clinical research. More validation with dentists and clinical data is needed to demonstrate its applicability in clinical practice.
- Hybrid architecture, GoogLeNet, Residual network (ResNet), Attention mechanism, Pre-processing, Hierarchical features
How to Cite
References
[1] Lahoud P, EzEldeen M, Beznik T, Willems H, Leite A, Van Gerven A, et al. Artificial intelligence for fast and accurate 3-dimensional tooth segmentation on cone-beam computed tomography. J Endod. 2021;47(5):827-35. [DOI: 10.1016/j.joen.2020.12.020]
[2] Shaheen E, Leite A, Alqahtani KA, Smolders A, Van Gerven A, Willems H, et al. A novel deep learning system for multi-class tooth segmentation and classification on cone beam computed tomography: a validation study. J Dent. 2021;115:103865. [DOI: 10.1016/j.jdent.2021.103865]
[3] Jang TJ, Kim KC, Cho HC, Seo JK. A fully automated method for 3D individual tooth identification and segmentation in dental CBCT. IEEE Trans Pattern Anal Mach Intell. 2022;44(10):6562-8. [DOI: 10.1109/TPAMI.2021.3086072]
[4] Li Q, Han L, Zhuang Y, Li J, Lin J. Automatic tooth root segmentation of cone beam computed tomography image sequences using U-Net and RNN. J Xray Sci Technol. 2020;28(5):905-922. [DOI: 10.3233/XST-200678]
[5] Hsu K, Yuh D, Lin S, Lyu P, Pan G, Zhuang Y, et al. Improving performance of deep learning models using 3.5D U-Net via majority voting for tooth segmentation on cone beam computed tomography. Sci Rep. 2022;12:19809. [DOI: 10.1038/s41598-022-23901-7]
[6] Wen X, Liu Z, Chu Y, Le M, Li L. MRCM-UCTransNet: automatic and accurate 3D tooth segmentation network from cone-beam CT images. Int J Imaging Syst Technol. 2024;34:e23139. [DOI: 10.1002/ima.23139]
[7] Wang Y, et al. Root canal treatment planning by automatic tooth and root canal segmentation in dental CBCT with deep multi-task feature learning. Med Image Anal. 2023;85:102750. [DOI: 10.1016/j.media.2023.102750]
[8] Hu F, Chen Z, Wu F. A novel difficult-to-segment samples focusing network for oral CBCT image segmentation. Sci Rep. 2024;14:5068. [DOI: 10.1038/s41598-024-55522-7]
[9] Polizzi A, Quinzi V, Ronsivalle V, Venezia P, Santonocito S, Lo Giudice A, et al. Tooth automatic segmentation from CBCT images: a systematic review. Clin Oral Investig. 2023;27(7):3363-78. [DOI: 10.1007/s00784-023-05048-5]
[10] Zheng Q, Gao Y, Zhou M, Li H, Lin J, Zhang W, et al. Semi- or fully automatic tooth segmentation in CBCT images: a review. PeerJ Comput Sci. 2024;10:e1994. [DOI: 10.7717/peerj-cs.1994]
[11] Chen X, Ma N, Xu T, Xu C. Deep learning-based tooth segmentation methods in medical imaging: a review. Proc Inst Mech Eng H. 2024;238(2):115-131. [DOI: 10.1177/09544119231217603]
[12] Zhou Y, Xu Y, Khalil B, Nalley A, Tarce M. An open deep learning-based framework and model for tooth instance segmentation in dental CBCT. Clin Oral Investig. 2025;29. [DOI: 10.1007/s00784-025-06578-w]
[13] Kot WY, Au Yeung SY, Leung Y, Leung P, Yang WF. Evolution of deep learning tooth segmentation from CT/CBCT images: a systematic review and meta-analysis. BMC Oral Health. 2025;25. [DOI: 10.1186/s12903-025-05984-6]
[14] Zhou W, Zhao D, Zhao B, Shen M, Liu X. A deep learning framework for automated dental segmentation and diagnostic report generation from cone-beam computed tomography. Head Face Med. 2025;21:80. [DOI: 10.1186/s13005-025-00555-0]
[15] Zakareya S, Izadkhah H, Karimpour J. A new deep-learning-based model for breast cancer diagnosis from medical images. Diagnostics (Basel). 2023;13(11):1944. [DOI: 10.3390/diagnostics13111944]
[16] Balraj K, Ramteke M, Mittal S, Bhargava R, Rathore AS. MADR-Net: multi-level attention dilated residual neural network for segmentation of medical images. Sci Rep. 2024;14. [DOI: 10.1038/s41598-024-63538-2]
- Abstract Viewed: 32 times
- 53264 Downloaded: 24 times