The subpubic angle and palpable pelvic parameters in sex estimation using machine learning algorithms in a Turkish...

Area of Science:

• Forensic Anthropology
• Medical Imaging
• Machine Learning

Background:

• Accurate sex estimation is crucial in forensic anthropology.
• Traditional methods rely on skeletal morphology, which can be challenging with fragmented remains.
• Computed tomography (CT) offers detailed pelvic measurements for advanced analysis.

Purpose of the Study:

• To estimate sex in the Turkish adult population using CT-derived pelvic morphometry.
• To evaluate the effectiveness of machine learning algorithms (MLAs) for sex estimation.
• To establish a reliable method for forensic applications.

Main Methods:

• Retrospective analysis of 301 adult CT scans (150 female, 151 male).
• Measurement of six pelvic parameters, including five palpable anatomical landmarks.
• Application of supervised (Logistic Regression, CART, MLP) and unsupervised (K-means) machine learning algorithms.

Main Results:

• Machine learning models achieved high accuracy in sex estimation.
• Multilayer Perceptron (MLP) demonstrated superior performance, reaching 100% accuracy for females and 96% for males.
• Logistic Regression (LR) achieved 94.7% overall accuracy, with most variables significantly contributing.

Conclusions:

• CT-based pelvic morphometry analyzed by MLAs provides a highly accurate and objective method for sex estimation.
• The developed models are applicable to the Turkish population, aiding forensic investigations.
• This approach is suitable for both living individuals and skeletal remains.