Related Experiment Video
Updated: Aug 5, 2026
02:42
Analysis of Craniomaxillofacial Malformations in Mice Using Three-dimensional Microcomputed Tomography
Published on: January 17, 2025
Radiographic legal age estimation based on third molar development using Machine-learning algorithms
Noura Alsufyani1, Sawsan Alowa2, Hend Alrasheed3
1Department of Oral Medicine and Diagnostic Sciences, College of Dentistry, King Saud University, Riyadh, Saudi Arabia.
Plos One
|July 30, 2026
Summary
Supervised machine learning (SML) accurately classifies individuals by age using third molar development, outperforming human experts in dental age estimation. This AI-driven approach offers a more precise method for forensic age assessment.
Area of Science:
- Forensic Science
- Radiology
- Artificial Intelligence
Background:
- Age assessment is crucial in forensic science, particularly for legal matters.
- Third molar development continues past the age of legal adulthood, making it a key indicator.
- Current radiographic dental age estimation methods are operator-dependent, introducing variability.
Purpose of the Study:
- To evaluate the performance of supervised machine learning (SML) for age classification.
- To classify individuals based on the 18-year legal age threshold using the third molar index (I3m).
- To compare the accuracy of SML models against expert human assessment.
Main Methods:
- Utilized panoramic radiographs (n=597) from individuals aged 13-26 years (52% male, 48% female).
- Employed Convolutional Neural Networks (CNNs) for third molar segmentation and various machine learning algorithms for classification.
- Implemented a 10-fold cross-validation approach and compared the best SML model against expert evaluations.
Main Results:
- Attention U-Net achieved superior third molar segmentation scores.
- K-nearest neighbors (KNN) demonstrated the best performance among classification algorithms.
- SML (KNN) achieved high sensitivity and specificity (e.g., males: 77.77% sensitivity, 96% specificity; females: 77.4% sensitivity, 86.7% specificity), outperforming expert scores in most metrics.
Conclusions:
- SML models accurately classify individuals above and below the legal age of 18, surpassing manual methods.
- AI, through pixel-level analysis, can identify subtle developmental cues imperceptible to human experts.
- This study suggests a potential shift from traditional dental age estimation indices towards AI-powered analysis.
