Sex estimation from biometric face photos for forensic purposes


Sezgin N., KARADAYI B.

Medicine, Science and the Law, 2022 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Publication Date: 2022
  • Doi Number: 10.1177/00258024221100898
  • Journal Name: Medicine, Science and the Law
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI), Scopus, Periodicals Index Online, Business Source Elite, Business Source Premier, CINAHL, EBSCO Legal Source, EMBASE, HeinOnline-Law Journal Library, MEDLINE
  • Keywords: anthropometric measurements, biometric photograph, discriminant analysis, sex estimation
  • Kütahya Health Sciences University Affiliated: Yes

Abstract

© The Author(s) 2022.Sex estimation is an important parameter in cases where individuals need to be identified in forensic cases. Biometric photographs are a form of a passport photo with specific dimensions and features established by the International Civil Aviation Organization, which are read and digitally stored in appropriate devices, are used in travel documents, and are of high quality (at least 600 dpi). This study aims to reveal anthropometric data for estimating sex in Turkish adult population from facial images conforming to biometric photography criteria. Within the scope of this study, biometric facial images of a total of 334 participants, 146 female and 188 male, between the ages of 20 and 79 were used. The photos were taken using a Nikon D5100 and flat front lighting from a distance of 1 m. ImageJ 1.50i software was used to process these images. Statistical analysis was performed using descriptive statistics and discriminant analysis tests. Among the 11 variables on sex estimation, the highest accuracy rate of 78.1% was obtained with the measurement between Gonion points. However, sex estimation could be made with an accuracy of 80.5% by including all age groups and all variables, and when age-specific data were used, it was observed that these accuracy rates increased significantly in all three age groups (84.6%, 89.2%, 85.2%, respectively). Therefore, we are suggesting that using age-specific data generated for estimation in different age groups. Consequently, it has been shown that successful sex estimation can be done by formulas derived from biometric facial images in this study.