Applicability of a pre-established set of facial proportions from frontal photographs in forensic age estimation of a Brazilian population.

Child pornography on the internet is one of the crimes with highest increasing incidence worldwide. In this specific type of crime, the victim's age has important legal implications. To overcome the lack of scientifically established protocols, European researchers explored and proposed a set of facial proportions obtained from images to estimate age. Within a validation scope, the present study aimed to test the applicability of morphological proportions from frontal photographs of the face to estimate age in a Brazilian population. Standardized frontal view photographs from 800 Brazilians (400 females and 400 males) homogeneously distributed in 4 age groups (6, 10, 14 and 18 years ± 30 days) were randomly selected from a Civilian Database of the Brazilian Federal Police. The selected images were analyzed with the aid of cephalometric landmarks. After landmarking the photographs, eighteen metric relations pre-established by Cattaneo et al. (C. Cattaneo, Z. Obertová, M. Ratnayake, L. Marasciuolo, J. Tutkuviene, P. Poppa, D. Gibelli, P. Gabriel, S. Ritz-Timme, Can facial proportions taken from images be of use for ageing in cases of suspected child pornography? A pilot study, Int. J. Leg. Med. 126 2012 139-144) were assessed and analyzed through parametric statistical tests and discriminant analysis. Two indexes had a moderate correlation with age, while ten had a weak correlation and six were not correlated with age. Eleven indexes were statistically different between sexes (p < 0.05). The discriminant analysis showed that 49.6% of the pooled sample was correctly classified into the respective age group. This percentage increased when each sex was analyzed separately: 50.2% for females and 53.5% for males. The set of facial proportions from frontal images proposed by Cattaneo et al. (2012) is related to facial development. As expected, the discriminant power of this approach did not reach optimal effectiveness in the Brazilian sample. In practice, this age estimation method must be adapted for more reliable application in Brazilians. [ABSTRACT FROM AUTHOR]