Facial soft tissue depth analysis and development for Greek population

Abstract

The process of identifying human remains involves using one of several methods, such as DNA analysis, dental records and fingerprint matching. These methods can only be used if there is antemortem data available for the assumed individuals. If antemortem data is not available or any of the identification methods cannot be performed, then an anthropologist can establish the biological profile of the skeletal remains to narrow down the possible identities. However, despite all efforts, the identification of the remains is still prone to potential failure. In such cases, a facial approximation of the person of interest is used as a last resort to reach individuals who might know the possible identity.

Facial approximation is a technique that involves constructing the facial muscles and applying a suitable facial soft tissue depth (FSTD) dataset. To establish a FSTD dataset, various methods are used. However, there have been limited validation studies conducted to date. Therefore, this study aims to compare the 2D and 3D measurement methods to investigate if they can be used interchangeably. Furthermore, this study presents the 3D FSTD datasets of an adult Greek population sample for the first time.

The CT scans of 50 male and 50 female adults (N=100; Mean age=58.9 years; SD=19.5) were obtained from Greece. For the 3D measurement method, the 3D skull and head models of each individual were created using Amira 6. Then, the facial depths were measured at the 22 landmarks. The facial depths were also measured directly from the CT slices by involving 15 landmarks in the 2D measurement method. Lastly, the facial approximations of six individuals were carried out for the qualitative assessment. The reproducibility of the 3D models and the technical error of the measurements were tested. Bilateral asymmetry across landmarks, the impact of age and sex on the FSTDs and FSTD variation across selected populations were investigated. The datasets of the 2D and 3D methods were compared. Finally, the FSTD datasets were evaluated qualitatively.

The results indicated that the 3D models were reproduced with high precision. Similarly, the technical error measurements showed high repeatability for most of the landmarks. The impact of age and sex on the FSTD was limited. Slight depth variations were obtained across the population samples. Statistically significant differences were obtained between the 2D and 3D FSTD datasets. The results of the qualitative assessment suggested that using two different FSTD datasets did not compromise facial recognition.

This current study validated that both 2D and 3D measurement methods provide similar FSTD values. Therefore, establishing FSTD databases using both methods would not compromise facial approximations. This study contributes to the population-specific FSTD dataset for adult individuals of Greek ancestry.