Gender Prediction From Foramen Magnum Using Machine Learning Algorithms

Abstract

Aim; Accurate and highly accurate postmortem identification of the individual is important for forensic sciences. The main element of identification is the determination of gender. The aim of this study is to predict gender with high accuracy using Machine learning (ML) algorithms with parameters of the foramen magnum obtained from Computed Tomography (CT) images. Method; The study was performed on CT images of 214 individuals aged 18-65 years. For. magnum length, width, circumference and area were measured on the images. The measurements were used in ML algorithms for gender prediction and performance values were recorded. Results; As a result of the study, gender prediction results with high accuracy up to 0.84 were obtained with ML algorithms. In addition, it was found that the for. magnum height parameter contributed the most to this high rate with the SHapley Additive exPlanations analyzer. Conclusion; As a result of our study, it was found that the parameters obtained from for. magnum can be used for sex prediction in the Turkish population by analyzing them with ML algorithm. In this respect, we think that it will contribute to other metric studies in the Turkish population.