Research on Automatic Diagnosis Model of Dental Radiographs Based on Deep Learning

DOI: https://doi.org/10.62517/jbdc.202601228

Author(s)

Zuoting Qin, Fang Fang*, Zhichen Xie

Affiliation(s)

School of Computer and Artificial Intelligence, Xiangnan University, Chenzhou, Hunan, China *Corresponding Author

Abstract

This study aims to address the challenges of low efficiency and high inter-observer variability in manual dental radiograph interpretation. An improved multi-scale residual network model is proposed for automatic detection and classification of common dental lesions. A dataset of 8,500 annotated periapical radiographs is constructed, and comprehensive preprocessing including noise reduction and contrast enhancement is performed. The model integrates channel attention mechanisms to capture subtle lesion features and adopts multi-scale feature fusion to handle lesions of different sizes. Experimental results demonstrate that the proposed model achieves 94.2% accuracy, 92.7% sensitivity, and 95.1% specificity, outperforming traditional convolutional neural network models and showing potential for clinical auxiliary diagnosis applications.

Keywords

Deep Learning; Dental Radiographs; Automatic Diagnosis; Residual Network; Medical Image Processing

References

[1] Xiang J, Miao J, Yang Y, et al. Auxiliary diagnosis system for dental caries based on deep learning[J]. Intelligent Computer and Applications, 2025, 15(3): 64-71. [2] Wu D, Wang P, Jing Y, et al. Construction and evaluation of an intelligent diagnostic model for temporomandibular joint osteoarthritis based on deep learning[J]. Journal of Practical Stomatology, 2025(4). [3] Guo W, Tang Y, Zhang X, et al. Comparison of CT image classification efficacy between deep learning models and radiologists in multidrug-resistant tuberculosis and drug-sensitive tuberculosis[J]. Journal of Clinical Pulmonary Medicine, 2026, 31(3): 407-411. [4] Wu W. Research on cardiovascular and cerebrovascular medical image diagnosis technology based on deep learning and multimodal fusion[J]. Information and Computer, 2025(21): 22-25. [5] Wang F, Wang P, Zhou L, et al. Construction and application of an intelligent image diagnosis platform based on deep learning[J]. China Digital Medicine, 2020, 15(1): 3. DOI: CNKI:SUN:YISZ.0.2020-01-006. [6] Wang K, Zhang G. Thyroid SPECT image diagnosis based on ResNet model[J]. Journal of Hebei University of Science and Technology, 2020, 41(3): 7. DOI: 10.7535/hbkd.2020yx03006. [7] Lin G, Zhang Q, Li Y, et al. Value of multi-label learning MRI knee joint sports injury detection model in auxiliary diagnosis[J]. Chinese Journal of Radiology, 2021. DOI: 10.3760/cma.j.cn112149-20201130-01266. [8] Wei B, Li Y, Zhang X, et al. Research progress of deep learning combined with radiomics in musculoskeletal diseases[J]. Chinese Journal of Magnetic Resonance Imaging, 2026, 17(3): 213-220. DOI: 10.12015/issn.1674-8034.2026.03.031. [9] Ma G, Yan C, Yang L, et al. Hepatic echinococcosis ultrasound image diagnosis method based on improved multi-scale deep residual network[J]. Journal of Northeast Normal University (Natural Science Edition), 2023, 55(1): 8. [10] Shen G, Lin H, Hu Y, et al. Comparative analysis of performance between traditional radiomics and deep learning in oral cancer image diagnosis[J]. Advances in Clinical Medicine, 2025, 15(9): 2034-2040. DOI: 10.12677/acm.2025.1592714. [11] Sun Y, Wang X, Xiao Y. Application progress of deep learning technology in lung cancer imaging diagnosis[J]. Shanghai Medical Imaging, 2021, 30(6): 30. [12] Anonymous. Research progress of deep learning in brain tumor MRI image diagnosis[J]. Journal of Frontiers of Computer Science and Technology, 2026. DOI: 10.3778/j.issn.1673-9418.2508070. [13] Wang J, Lin Y, Xiong J, et al. Application of deep learning-based spontaneous intracerebral hemorrhage CT image segmentation algorithm in accurate calculation of lesion volume[J]. Chinese Journal of Radiology, 2019, 53(11): 941-945. DOI: 10.3760/cma.j.issn.1005-1201.2019.11.003. [14] Zhao D. Research progress of deep learning and medical image analysis[J]. Electronic Science and Technology, 2018, 11(4): 4. [15] Hang Y, Gao H. Application of artificial intelligence technology in pulmonary nodule imaging diagnosis: progress, challenges and prospects[J]. Journal of Clinical Personalized Medicine, 2024, 3(4): 1896-1902. DOI: 10.12677/jcpm.2024.34266. [16] Hou D. Application of deep learning in pulmonary tuberculosis image diagnosis[J]. Chinese Journal of Antituberculosis, 2022, 44(1): 4. DOI: 10.19982/j.issn.1000-6621.20210537.