Considering the high demand for the identification of unknown remains in South Africa, a need exists to establish reliable facial approximation techniques that will take into account sex and age and, most importantly, be useful within the South African context. This study aimed to provide accurate statistical models for predicting nasal soft-tissue shape from information about the underlying skull subtract among a South African sample.

The database containing 200 cone-beam computer tomography (CBCT) scans (100 black South Africans and 100 white South Africans). The acquisition and extraction of the 3D relevant anatomical structures (hard- and soft-tissue) were performed by an automated three-dimensional (3D) method based on an automatic dense landmarking procedure using MeVisLab (c) v. 2.7.1 software. An evaluation of shape differences attributed to known factors (ancestry, sex, size, and age) was performed using geometric morphometric and statistical models of prediction were created using a Projection onto Latent Structures Regression (PLSR) algorithm. The accuracy of the estimated soft-tissue nose was evaluated in terms of metric deviations on training and un-trained datasets.

Our findings demonstrated the influence of factors (sex, aging, and allometry) on the variability of the hard- and soft-tissue among two South African population groups. This research provides accurate statistical models optimized by including additional information such as ancestry, sex, and age. When using the landmark-to landmark distances, the prediction errors ranged between 1.769 mm and 2.164 mm for black South Africans at the tip of the nose and the alae, while they ranged from 2.068 mm to 2.175 mm for the white subsample. The prediction errors on un-trained data were slightly larger, ranging between 2.139 mm and 2.833 mm for the black South African sample at the tip of the nose and the alae and ranging from 2.575 mm to 2.859 mm for the white South African sample.

This research demonstrates the utilization of an automated 3Dmethod based on an automatic landmarking method as a convenient prerequisite for providing a valid and reliable nose prediction model that meets population-specific standards for South Africans. (C) 2020 Elsevier B.V. All rights reserved.