Gender assessment from upper face height and width in skull using cone beam computed tomography

Craniofacial anthropometry, which is a branch of anthropometry, includes head and face measurements on living, cadaver and radiological samples. These measurements are important data for surgical and forensic applications. In the present study, the purpose was to evaluate gender in adults by examining Cone Beam Computed Tomography (CBCT) images from upper face width and upper face height measurements in the skull. A total of 2 parameters (i.e., upper face width and upper face height) were evaluated on CBCT images of 52 female and 52 male adults in the present study. With the help of parameter measurements, the success of various machine learning algorithms, Logistic Regression (LR), Support Vector Machines (SVM), Artificial Neural Networks (NN), Classification and Regression Trees (CART) in gender discrimination was examined. Although the mean age of men and women between the ages of 18-70 included in the study was not found to be significantly different (p=0.794), the mean of upper face width and upper face height measurement was significantly higher in males (p=0.001). It was observed that the algorithms used in gender determination were the CART model with the highest overall diagnostic success rate of 89.4%. In the present study, it was concluded that gender can be evaluated with higher diagnostic success than the upper face width and upper face height measurements in the skull. Also, the selection of the algorithm that reveals the purpose of the study is important for correct classification success because the success criteria of the algorithm in gender estimation will be different.