Morfometria geométrica em fotografias faciais frontais para avaliação do crescimento e diagnóstico de idade

• Autor(a) principal: Damascena, Nicole Prata
• Data de Publicação: 2023
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da UFS
• Texto Completo: https://ri.ufs.br/jspui/handle/riufs/23381
• topic: Análise facial; Análise morfológica; Antropologia forense; Estimativa de idade; Diagnóstico de idade; Estimativa de sexo; Morfometria geométrica; Facial analysis; Morphological analysis; Forensic anthropology; Age estimation; Age diagnosis; Sex estimation; Geometric morphometry

The study of facial growth with a focus on age estimation and diagnosis is of great importance to medical, dental, and forensic sciences. However, there is currently no scientifically validated method to support forensic results based solely on images, which are often the only available material for analysis. The aim of this study was to evaluate the applicability of Geometric Morphometrics in detecting facial growth patterns in Brazilian children, adolescents, and young adults of both sexes, for the purpose of age estimation in forensic contexts. Automated facial analysis was performed on 4,000 frontal photographs of Brazilian individuals from the civil image database of the Federal Police, equally divided into age groups of 6, 10, 14, and 18 years, and both female and male sexes. The analyses were conducted using R software (version 3.6), with a significance level of 5%. The facial pattern of the sample was modeled using the statistical package geomorph, and the application of Geometric Morphometrics as an age estimation method was based on the Procrustes paradigm. The performance of the method in terms of accuracy, sensitivity, and specificity was tested a priori using a Multinomial Logistic Regression model. The observed growth pattern showed a vertical increase in the face in both sexes over the years, with a slower and more pronounced growth in males, particularly in the lower third of the face. The Geometric Morphometrics methodology showed interesting age discrimination values, achieving an accuracy rate of approximately 70% (69.3%, 95% CI 72.7 - 76.6) in age prediction. The best performance of the method was observed for the age of 6 years (Sensitivity: 87.3%, 95% CI 85.1 - 89.3; Specificity: 95.6%, 95% CI 94.1 - 96.8), especially in the presence of photographs of female children. For the other age groups, Geometric Morphometrics exhibited better performance in age estimation among male individuals. Thus, the study contributed to establishing population growth patterns using facial Geometric Morphometrics, indicating it as a promising method for age estimation in forensic contexts.