Tooth shape and sex estimation: A 3D geometric morphometric landmark-based comparative analysis of artificial neural networks, support vector machines, and random forest models

Introduction: Sex estimation from dental morphology is crucial in forensic identification; AI and 3D morphometrics may improve accuracy. Methods: 3D scans of dental casts from 120 individuals (60M/60F). Nine tooth types landmarked, Procrustes superimposition and PCA performed. PCA features used to train SVM, ANN, and RF. Performance assessed via 5-fold cross-validation: accuracy, precision, recall, F1, AUC. Results: Random Forest performed best (up to 97.95% accuracy for mandibular second premolars). SVM showed moderate accuracy (70–88%). ANN performed poorest (58–70%) with lower female classification recall. Mandibular premolars were most dimorphic. Conclusions: RF is the most robust model for sex estimation using 3D dental landmarks. Traditional ML approaches outperformed ANN here; hybrid models merit future exploration.