Akademik Veri Yönetim Sistemi
ASERC 2nd International Conference On Health, Engineering, Architecture And Mathematics, İstanbul, Türkiye, 6 - 08 Haziran 2025, ss.359-404, (Tam Metin Bildiri)
ABSTRACT
Aim
This study aims to systematically review machine learning (ML) methods used for sex and age
estimation in forensic dentistry, with a specific focus on the contribution of explainable artificial
intelligence (AI) technologies.
Methodology
The study followed PRISMA guidelines for systematic reviews. A comprehensive literature
search was conducted in PubMed, Scopus, Web of Science, Google Scholar, and IEEE Xplore
for studies published between January 2021 and March 2025. Studies included had to use ML
or deep learning (DL) methods for sex or age estimation with a minimum sample size of 50.
Extracted parameters included algorithm type, imaging modality, sample size, and performance
metrics such as accuracy, sensitivity, specificity, and F1-score.
Result
Deep learning models, particularly Convolutional Neural Networks (CNNs) and Vision
Transformers applied to 3D imaging data, demonstrated the highest performance. Hybrid
models combining imaging and morphometric features improved model generalizability.
Studies using multicenter datasets enhanced external validity. AI tools such as heatmaps and
feature attribution methods significantly improved model interpretability and addressed ethical
and forensic transparency concerns.
Discussion
ML and AI approaches reduce subjectivity in forensic assessment and provide more consistent
and defensible biological profile estimations. However, variability in methodology, sample
heterogeneity, and limited transparency in some models pose challenges for standardization. AI
plays a crucial role by providing insight into decision-making processes, thus enhancing legal
credibility.
Conclusion
AI-enhanced ML models represent a significant advancement in forensic dentistry by
increasing diagnostic precision and legal robustness. Standardized workflows, extensive
external validation, and ethical oversight frameworks are essential for broader adoption.
Keywords: Forensic Dentistry, Machine Learning, Deep Learning, Sex Estimation, Age
Estimation