Related Experiment Video
Updated: May 2, 2026
07:57
Accuracy in Dental Medicine, A New Way to Measure Trueness and Precision
Published on: April 29, 2014
13.4K
Evaluation of Caries Detection on Bitewing Radiographs: A Comparative Analysis of the Improved Deep Learning Model
Baturalp Ayhan1, Enes Ayan2, Gökhan Karadağ3
1Department of Restorative Dentistry, Faculty of Dentistry, Bursa Uludag University, Bursa, Turkey.
Summary
A new deep learning model, YOLOv9c, effectively detects dental caries in bitewing radiographs. This advanced artificial intelligence tool demonstrated superior performance compared to other models and even dentists, aiding in accurate diagnosis.
Area of Science:
- Artificial Intelligence in Dentistry
- Medical Imaging Analysis
- Deep Learning for Diagnostics
Background:
- Deep learning models are increasingly used for detecting dental caries in radiographs.
- Comparative analysis of modern deep learning models for caries detection requires further investigation.
Purpose of the Study:
- To evaluate 11 You Only Look Once (YOLO) object detection models for identifying enamel and dentin caries.
- To refine the YOLOv9c model's architecture for improved detection performance and efficiency.
- To compare the YOLOv9c model's performance against dentists in detecting caries.
Main Methods:
- Eleven You Only Look Once (YOLO) object detection models were assessed for automatic caries identification.
- The YOLOv9c model underwent backbone architecture refinement to reduce size and computational load.
- The enhanced YOLOv9c model was evaluated against six dentists using identical bitewing radiograph datasets.
Main Results:
- The refined YOLOv9c model achieved superior performance metrics, including recall (0.727), precision (0.651), specificity (0.726), F1-score (0.687), and Youden index (0.453).
- The YOLOv9c model's recall and F1-score surpassed the performance of the six participating dentists.
- The optimized YOLOv9c model demonstrated high accuracy in detecting both enamel and dentin caries.
Conclusions:
- The YOLOv9c model is highly effective for detecting enamel and dentin caries, outperforming existing models and dentists.
- This advanced deep learning model can serve as a valuable tool to enhance dentists' diagnostic capabilities.
- The YOLOv9c model shows significant potential for clinical application, supporting accurate and efficient caries detection to improve patient outcomes.
