Related Experiment Video
Updated: Jun 7, 2025
08:07
A Morphometric and Cellular Analysis Method for the Murine Mandibular Condyle
Published on: January 11, 2018
8.2K
Sex prediction through machine learning utilizing mandibular condyles, coronoid processes, and sigmoid notches
Isabela Bittencourt Basso1, Pedro Felipe de Jesus Freitas2, Aline Xavier Ferraz3,4
1Postgraduate Program in Dentistry, Pontifícia Universidade Católica do Paraná, Curitiba, Brazil.
Plos One
|November 15, 2024
Summary
Mandible bone morphology, including the coronoid process, condyle, and sigmoid notch, can predict sex using machine learning. The distance between the condyle and gnathion (Co-Gn) was a key predictor in this study.
Area of Science:
- Forensic Anthropology
- Biometrics
- Medical Imaging Analysis
Background:
- Mandibular morphology is crucial for sex determination in anthropological and forensic contexts.
- Previous studies have explored various skeletal markers for sex prediction, but advanced computational methods offer new possibilities.
Purpose of the Study:
- To evaluate the efficacy of coronoid process, condyle, and sigmoid notch morphology in sex prediction using supervised machine learning algorithms.
- To identify the most significant morphometric variables for sex determination from cephalometric radiographs.
Main Methods:
- Analysis of cephalometric radiographs from 410 dental records, focusing on coronoid process, condyle, sigmoid notch, and Co-Gn distance.
- Application of seven machine learning algorithms: Decision Tree, Gradient Boosting Classifier, K-Nearest Neighbors (KNN), Logistic Regression, Multilayer Perceptron Classifier, Random Forest Classifier, and Support Vector Machine (SVM).
- Validation using 5-fold cross-validation, with performance metrics including AUC, accuracy, recall, precision, and F1 Score.
Main Results:
- The Co-Gn distance emerged as the most influential variable for sex prediction across multiple algorithms.
- Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Gradient Boosting Classifier demonstrated the highest performance (AUC and precision) in both cross-validation and testing.
- Model performance varied, with AUC ranging from 0.66 to 0.82 on test data and precision from 0.68 to 0.83.
Conclusions:
- Morphological characteristics of the mandible, particularly the Co-Gn distance, combined with machine learning, show promise for sex prediction.
- The study highlights the potential of SVM, KNN, and Gradient Boosting Classifier for this application.
- Findings should be interpreted cautiously due to study limitations, emphasizing the need for further research.
