Automated Classification of Dental Caries in Bitewing Radiographs Using Machine Learning and the ICCMS Framework
Bibliographic record
Abstract
Background: Dental caries is considered a public health issue, with early detection being crucial for effective management. Traditional diagnostic methods, including visual examination and bitewing radiographs, are prone to interpretation variability. Artificial intelligence (AI), particularly deep learning (DL), has shown promise in improving diagnostic accuracy. This study evaluates the YOLOv11 model for dental caries detection and segmentation in bitewing radiographs, using the standardized International Caries Classification and Management System (ICCMS) framework. Methods: A dataset of 730 bitewing radiographs, containing 1115 annotated carious lesions, was used for training and validation. Annotation was performed by experienced dentists using the Roboflow platform. To evaluate annotation consistency, a subset of 10 images was independently annotated by both dentists. Agreement was assessed using Intersection over Union (IoU) and Dice similarity coefficient (DSC). The YOLOv11 model was trained for 50 epochs with data augmentation techniques. Performance was assessed using precision (P), recall (R), and mean average precision at 50% IoU (mAP50). Results: The reliability analysis showed strong agreement, with an average interrater IoU of 0.82 and DSC of 0.85, and intrarater IoU of 0.84 and DSC of 0.87 across the 10 images. The YOLOv11 model excelled in detecting and segmenting advanced carious lesions, achieving high mAP50 values of 0.74 and 0.80 for RB4 + RC5 and RC6 classes, respectively. However, it showed moderate performance for early‐stage lesions (RA1 + RA2 and RA3), with mAP50 scores of 0.61 and 0.52, respectively. This disparity highlights areas for potential enhancement through additional data augmentation and model fine‐tuning. Conclusion: The YOLOv11 model is highly effective in identifying dental caries, especially advanced lesions, but struggles with detecting early stages of caries. AI enhancements could improve diagnostic accuracy, enable better early interventions and improve patient outcomes. The research supports incorporating AI technologies into dental radiographic evaluations to improve diagnostics and clinical results.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".