Anthropometric measurements of human faces generated by artificial intelligence

Abstract

Objectives: Artificial intelligence (AI) systems are capable of detecting human faces from two-dimensional images and generating highly realistic facial representations that do not correspond to any real individuals. This study aims to quantitatively assess the anthropometric features of AI-generated virtual faces and compare these measurements with established facial anthropometric data across different human populations. Methods: A total of 150 virtual faces (75 male, 75 female) were generated by an artificial intelligence system trained on the CelebAMask-HQ dataset, which consists of 30,000 high-resolution facial images of celebrities. Anthropometric distances between defined facial landmarks were measured using custom-developed software. The obtained measurements were statistically compared with anthropometric reference data from various populations reported in the literature. Statistical analysis was performed using the One-Sample t-test to assess deviations from known population means, and the Chi-square Goodness-of-Fit (χ²) test to evaluate distribution conformity. A significance level of p<0.05 was used for all analyses. Results: Several periorbital measurements of the AI-generated male virtual faces demonstrated greater similarity to anthropometric data from East Asian populations. Additionally, morphologic face height and nasal height values in male virtual faces were most closely aligned with those reported for Thai, Azeri, and Bulgarian populations. In female virtual faces, the circumference around the eyes was found to be comparable to that of Turkish females. Although certain facial features—particularly nasal and ocular parameters—showed resemblance to those of specific ethnic groups, the overall facial composition of both male and female virtual faces did not consistently correspond to any single racial or ethnic population. Conclusion: AI-generated virtual faces offer a novel and efficient alternative for establishing standardized anthropometric datasets representative of various ethnic groups. Instead of collecting data from large populations, artificial intelligence can generate virtual facial models based on existing datasets, from which reliable anthropometric measurements can be obtained. These virtual datasets can enhance diversity representation while minimizing racial bias and ethical concerns. Consequently, the anthropometric data derived from AI-generated faces may serve as a standardized reference across populations, supporting applications in forensic science, aesthetic surgery, ergonomics, and facial recognition technologies.

Keywords

• aesthetic plastic surgery
• artificial intelligence in health
• facial anthropometry
• generative adversarial networks
• photogrammetry

References

1. Choi HI, Jung SK, Baek SH, Lim WH, Ahn SJ, Yang IH, Kim TW. Artificial intelligent model with neural network machine learning for the diagnosis of orthognathic surgery. J Craniofac Surg 2019;30:1986–9.
2. Lu H, Li Y, Chen M, Kim H, Serikawa S. Brain intelligence: go beyond artificial intelligence. Mobile Networks and Applications 2018;23:368–75.
3. Zhang L, Tan J, Han D, Zhu H. From machine learning to deep learning: progress in machine intelligence for rational drug discovery. Drug Discov Today 2017;22:1680–5.
4. Tan C, Sun F, Kong T, Zhang W, Yang C, Liu C. A survey on deep transfer learning. In International conference on artificial neural networks. Springer, Cham; 2018. pp. 270–9.
5. Dilsizian SE, Siegel EL. Artificial intelligence in medicine and cardiac imaging: harnessing big data and advanced computing to provide personalized medical diagnosis and treatment. Curr Cardiol Rep 2014;16:441.
6. Hamet P, Tremblay J. Artificial intelligence in medicine. Metabolism 2017;69S:36–40.
7. Ramesh AN, Kambhampati C, Monson JR, Drew PJ. Artificial intelligence in medicine. Ann Roy Coll Surg 2004;86:334–8.
8. Vegter F, Hage J. Facial anthropometry in cleft patients: a historical appraisal. Cleft Palate Craniofac J 2001;38:577–81.

9. Sforza C, Ferrario VF. Soft-tissue facial anthropometry in three dimensions: from anatomical landmarks to digital morphology in research, clinics and forensic anthropology. J Anthropol Sc 2006;84:97–124.
10. Dere F, Oğuz Ö. Artistik anatomi. Adana: Nobel Tıp Kitabevi;1996. p. 110–4.
11. Yaşlı H, Bulut Ö. Morfolojik ve antropometrik yöntemlerin yüz karşılaştırma işlemlerinde uygulanması. Adli Bilimler Dergisi 2008;7:7–16.
12. He ZJ, Jian XC, Wu XS, Gao X, Zhou SH, Zhong XH. Anthropometric measurement and analysis of the external nasal soft tissue in 119 young Han Chinese adults. J Craniofac Surg 2009;20:1347–51.
13. Husein OF, Sepehr A, Garg R, Sina-Khadiv M, Gattu S, Waltzman J, Edward CWu, Mason Shieh, Gregory MH, Galle SE. Anthropometric and aesthetic analysis of the Indian American woman’s face. J Plast Reconstr Aesthet Surg 2010;63:1825–31.
14. Mishima K, Mori Y, Yamada T, Sugahara T. Anthropometric analysis of the nose in the Japanese. Cells Tissues Organs 2002;170:198–206.
15. Gulsen A, Candan O, Aslan BI, Uner O, Yavuzer R. The relationship between craniofacial structures and the nose in Anatolian Turkish adults: a cephalometric evaluation. Am J Orthod Dentofacial Orthop 2006;130:131.e15–25.
16. Dalal AB, Phadke SR. Morphometric analysis of face in dysmorphology. Comput Methods Programs Biomed 2007;85:165–72.
17. Yang YH, Wang B, Ding Y, Shi YW, Wang XG. Facial anthropometric proportion of Chinese Han nationality. J Craniofac Surg 2019;30:1601–4.
18. Leong SC, Eccles R. Race and ethnicity in nasal plastic surgery: a need for science. Facial Plast Surg 2010;26:63–8.
19. This Person Does Not Exist Website. [Internet]. [Retrieved on January 01, 2020]. Available from: https://www.thispersondoesnotexist.com/
20. Karras T, Aila T, Laine S, Lehtinen J. Progressive growing of GANs for improved quality, stability, and variation. Neural and Evolutionary Computing 2017;arXiv:1710.10196v3.
21. Dawei W, Guozheng Q, Mingli Z, Farkas LG. Differences in horizontal, neoclassical facial canons in Chinese (Han) and North American Caucasian populations. Aesthetic Plast Surg 1997;21:265–29.
22. Kim YC, Kwon JG, Kim SC, Huh CH, Kim HJ, Oh TS, Koh KS, Choi JW, Jeong WS. Comparison of periorbital anthropometry between beauty pageant contestants and ordinary young women with Korean ethnicity: a three-dimensional photogrammetric analysis. Aesthetic Plast Surg 2018;42:479–90.
23. Farkas LG, Katic MJ, Forrest CR, Alt KW, Bagic I, Baltadjiev G, Cunha E, Cvicelová M, Davies S, Erasmus I, Gillett-Netting R, Hajnis K, Kemkes-Grottenthaler A, Khomyakova I, Kumi A, Kgamphe JS, Kayo-daigo N, Le T, Malinowski A, Negasheva M, Manolis S, Ogetürk M, Parvizrad R, Rösing F, Sahu P, Sforza C, Sivkov S, Sultanova N, Tomazo-Ravnik T, Tóth G, Uzun A, Yahia E. International anthropometric study of facial morphology in various ethnic groups/races. J Craniofac Surg 2005;16:615–46.
24. Evereklioglu C, Doganay S, Er H, Gunduz A, Tercan M, Balat A, Cumurcu T. Craniofacial antropometry in a Turkish population. Cleft Palate Craniofac J 2002;39:208–18.
25. Farkas LG, Katic MJ, Forrest CR. Comparison of craniofacial measurements of young adult African-American and North American white males and females. Ann Plast Surg 2007;59:692–8.