Research on Intelligent Detection Algorithm of Dental Diseases Based on YOLO Algorithm

DOI: https://doi.org/10.62517/jmhs.202605121

Author(s)

Jiarui Luo, Zhengju Guo, Ting Huang, Shengni Fu, Yu Pu, Guangyan Wang*

Affiliation(s)

School of Information Engineering, Tianjin University of Commerce, Tianjin, China *Corresponding Author

Abstract

Amid the digital transformation of stomatology, deep learning-driven intelligent dental disease diagnosis is a research hotspot, yet traditional YOLOv5 has high small lesion miss detection and insufficient feature fusion in oral endoscopic image analysis. This study proposes an improved YOLOv5 algorithm, integrating CBAM and BiFPN for enhanced tiny lesion feature extraction and structured pruning with INT8 quantization for model lightweighting. We built a high-quality annotated dataset with tertiary hospital clinical images and MICCAI2023 public dataset (split 7:2:1 for training, validation and testing), and developed a PyQt5-based cross-platform system for clinical chairside use. The improved model achieved 89.5% precision, 88.1% recall, 88.8% F1-score, 92.3% mAP@0.5 (4.7 percentage points higher than original YOLOv5s) and 78.6% mAP@0.5:0.95 on the independent test set. Ablation experiments confirmed CBAM and BiFPN significantly boosted detection performance with only slight parameter growth. This algorithm resolves YOLOv5’s application limitations, providing an accurate and efficient intelligent auxiliary diagnosis scheme for clinical chairside scenarios, improving early dental disease diagnosis in primary care and advancing the digital and intelligent development of oral healthcare to support the Healthy China strategy.

Keywords

YOLOv5; Dental Disease Detection; Oral Endoscopy; CBAM Attention Mechanism; Bifpn Feature Fusion

References

[1]Li, Y., Zhang, H., & Wang, X. Artificial intelligence tools in dentistry: A systematic review on their application and outcomes. Healthcare, 2025, 13(10), 1452. [2]Kandaswamy, D., Priya, S., & Kumar, P. Developing an AI-based application for caries index detection on intraoral photographs. Computers in Biology and Medicine, 2024, 174, 107968. [3]Abidin, N. Z., Rahim, N. A., & Yusoff, N. Evaluating YOLO for dental caries diagnosis: a systematic review and meta-analysis. Journal of Dentistry, 2025, 154, 104542. [4]Shaji, A., Thomas, S., & Mathew, J. Visual diagnostics of dental caries through deep learning of non-standardised photographs using a hybrid YOLO ensemble and transfer learning model. International Journal of Environmental Research and Public Health, 2023, 20(7), 5351. [5]Tan, M., & Le, Q. V. EfficientDet: Scalable and efficient object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(11), 2020, 3713-3727. [6]Dogan, S., Kaya, M., & Yildirim, O. Diagnosis and classification of dental caries with deep learning methods in dental radiography. International Journal of Theoretical and Applied Physics, 2025, 17(2), 12-20. [7]Alom, M. Z., Hasan, M., & Yakopcic, C. U-NET model based on CBAM attention mechanism for coronary angiography segmentation. Current Medical Imaging, 2024, 20(10), 1089-1100. [8]SHORTEN C, KHOSHGOFTAAR T M. A survey on Image Data Augmentation for Deep Learning. Journal of Big Data, 2019, 6: 60. [9]WOO S, PARK J, LEE J Y, et al. CBAM: Convolutional Block Attention Module[C]//Proceedings of the European Conference on Computer Vision (ECCV). 2018: 3-19. [10]PADILLA R, NETTO S L, DA SILVA E A B. A Survey on Performance Metrics for Object-Detection Algorithms. 2020 International Conference on Systems, Signals and Image Processing (IWSSIP). 2020: 237-242. [11]Ramirez-Pedraza A, Salazar-Colores S, Cardenas-Valle C, et al. Deep Learning in Oral Hygiene: Automated Dental Plaque Detection via YOLO Frameworks and Quantification Using the O’Leary Index. Diagnostics. 2025, 15(2): 231. [12]Evaluation of the Performance of a YOLOv10-Based Deep Learning Model for Tooth Detection and Numbering on Panoramic Radiographs of Patients in the Mixed Dentition Period. Diagnostics. 2025, 15(4): 405. [13]N. N W, M. J, X. F. Simulation-Based Robustness Evaluation of Coordination under Stochastic Demand. International Journal of Simulation Modeling,2025,24(4):730-741. [14]Weiming Z, Patricia B, Allison L, et al. Real-World Positive Predictive Value and False Positive Rates of Laboratory-Based HIV Antigen/Antibody Testing by Age and Sex — United States, 2019–2024.JAIDS Journal of Acquired Immune Deficiency Syndromes,2025. [15]Zhou J, Zou J, Qiu Y, et al. Polarization image fusion via analytical attention heads: A multi-scale feature integration framework. Optics and Lasers in Engineering, 2026, 2011, 09628-109628. [16]Redmon J, Farhadi A. YOLOv3: An Incremental Improvement. arXiv preprint arXiv:1804.02767, 2018. [17]Lin T Y, Dollár P, Girshick R, et al. Feature pyramid networks for object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 2117-2125. [18]Li A, Qiu B, Goettsch M, et al. Convolutional neural networks combined with classification algorithms for the diagnosis of periodontitis. Journal of Clinical Periodontology, 2023, 50(5): 591-603. [19]Ozsunkar P S, Özen D Ç, Abdelkarim A Z, et al. Detecting white spot lesions on post-orthodontic oral photographs using deep learning based on the YOLOv5x algorithm: a pilot study. BMC Oral Health, 2025, 25(1): 1128.