Using Artificial Intelligence to Quantify Sexual Dimorphism in Aesthetic Faces: Analysis of 100 Facial Points in 42 Caucasian Celebrities.

Abstract

Sexual dimorphism has been studied in the faces of average populations and worldwide celebrities; however, a focused analysis of attractive Caucasian faces has not been conducted. The study harnesses the power of artificial intelligence (AI) to efficiently analyze these facial patterns in attractive Caucasian male and female celebrities. Twenty-one male and 21 female Caucasian celebrities were selected based on popular editorial rankings, modeling agencies, and casting directors from 2017 to 2022. Frontal photographs of celebrities aged 23 to 42 without facial animation were selected. One hundred facial landmarks were identified using semi-automatic image analysis software consisting of modified Apple Vision (Cupertino, CA) machine-learning algorithms with additional custom landmarks. Measurements were converted to absolute distances by fixing subjects' white-to-white corneal diameters to the validated average in Caucasians. Attractive females had significantly greater upper and middle facial proportions, more uniformly divided facial thirds, and greater canthal tilt compared with males. Attractive males had significantly greater facial height, bizygomatic and bigonial widths, medial and total brow lengths, and alar width than females. The golden ratio (1.618) was observed in the ratio of facial height to bigonial width in females (1.613), and attractive males closely approximated that ratio (1.566). There were no significant differences in interpupillary distances, eyebrow angles, or horizontal palpebral fissure lengths. No faces in either sex exhibited scleral show. The study is the first to utilize AI in quantifying key sexual dimorphisms among Caucasian celebrity faces. Identifying these contemporary patterns may provide valuable considerations in planning facial aesthetic and gender affirmation surgery.