Moscow University
Anthropology
Bulletin
Anthropology
Bulletin
1) Lomonosov Moscow State University, Anuchin Research Institute and Museum of Anthropology, Mokhovaya st., 11, Moscow, 125009, Russia; 2) Department of Biomedical Research, Main Department of Criminalistics, Investigation Committee of Russia, Moscow
Veselkova D.V., e-mail: daria.veselkova@yandex.ru; Goncharova Natalia N., Ph.D., ORCID ID: 0000-0001-8504-1175, e-mail: 1455008@gmail.com; Abramov A.S., e-mail: idenfac@gmail.com
A large number of methods are applied for describing human face. Using a scheme, similar to constitutional schemes, seems convenient for purposes of personal identification; however, there are almost no such schemes for the face. In some constitutional schemes, facial characteristics are included amongst other traits. However, the relationship between parameters of the body and face is low; therefore, it makes sense to consider the variability of the facial characteristics separately. To take into account existent variability and to reflect the continuity of facial features’ variation we used the approach formulated by V.E. Deryabin. It involves prior identification of variability pattern and a subsequent creation of a constitutional scheme using methods of multivariate analysis. We applied this approach to create a typological scheme based on measurements taken from photos, because working with photographic material is an important part of forensic experts’ practice. In this research, we used passport and anthropological full-face photos of 680 individuals in total. The measurements were performed in the «Bars Poisk» program using habitoscopic methodology and included 8 face parameters. Statistical data analysis and development of typology were performed in «Statistica 8» program using the principal component analysis method. According to the results of correlation analysis, two groups of parameters were distinguished, with higher level of association within than between these groups. However, even within these groups the level of parameters’ correlation was intermediate or low. Comparison of correlation matrixes obtained from measurement of photos with the literature data shows that the degree of relationship between features measured on a photo is higher than between features measured on a living human face and is comparable with the degree of relationship between the features of the facial skeleton. This suggests similar pattern of variability of skull features and an opportunity to compare face and skull images within the developed typology. Results of the component analysis produced three principal components which cumulatively described 68,2% and 69,1% of variability for males and females respectively. The first principal component takes into account total face size, dividing the whole sample within the micro-/macrosomia coordinate; the second principal component accounts for height-width proportions of the face and the third one – for the ratio between the sizes of the mouth and nose, i.e. central part of the face. A typological scheme built in coordinates of the second and third principal components reflects the variability of both general face shape and shape of its central part. The low correlations between the face features suggest an existence of a large number of their combinations. The approach, which was used for creating the face typology, allows to consider these combinations and objectively describe the observed pattern of variability. The resulting typology is easily formalized, which allows to objectify human face description based on a set of measurements.
anthropology, morphology, face typology, principal components, craniofacial identification, correlation analysis
Цит.: Veselkova D.V., Goncharova N.N., Abramov A.S. A morphological typology of faces. An experience of application of the principal component analysis in craniofacial identification // Moscow University Anthropology Bulletin (Vestnik Moskovskogo Universiteta. Seria XXIII. Antropologia), 2016; 3/2016; с. 4-16
Download text