Key points are not available for this paper at this time.
Abstract Background Identification is very difficult in situations where body integrity is compromised, such as explosions, war, airplane accidents and natural disasters. In such cases, sex estimation greatly facilitates identification. The aim of this study is to perform reliable and highly accurate sex estimation using machine learning algorithms with parameters obtained from the distal end of the humerus. The study was performed retrospectively on Computed Tomography Angiography images of 310 individuals aged 18–65 years. The distance between lateral epicondyle and medialepicondyle, width of the articular surface, width of the olecranon fossa, width of the radial fossa, width of the coronoid fossa, depth of the radial fossa, depth of the coronoid fossa, width of the capitulum, width of the trochlea, length of the capitulum humeri, length of the trochlea, the distance between the proximal point of the olecranon fossa and distal point of the trochlea, distance between the most distal points of trochlea humeri (DDT) were measured. Machine learning (ML) algorithms were analyzed with the data obtained. Results As a result of our study, a Sex estimation rate between 0.87 and 0.97 was obtained using ML algorithms with the parameters obtained from the distal end of the humerus. Using the SHAP, it was found that the DDT parameter made the highest contribution to sex estimation among the parameters used. Conclusion As a result of the study, it was found that the parameters obtained from the distal end of the humerus provided high accuracy in sex estimation using ML algorithms. In this respect, we believe that it will guide forensic studies.
Çayabatmaz et al. (Wed,) studied this question.