Predictors of female age, health and attractiveness perception from skin feature analysis of digital portraits in five ethnic groups

Abstract

AbstractObjectiveResearch indicates the impact of skin colour, tone evenness and surface topography on ratings of age, health and attractiveness in women. In addition to subjective assessments, these effects have been quantified with objective measures derived from skin image analysis. Signs of skin ageing may manifest differently across ethnic groups. However, comparisons have been limited to research with two ethnic groups, preventing conclusions about an ethnicity‐specific ranking of skin ageing signs.MethodsWe report results from a multi‐ethnic and multi‐centre study in which faces of women (n = 180; aged 20–69 years) from five ethnic groups were imaged. Facial images were rated for age, health and attractiveness by members of the same ethnic group (each n = 120). Digital image analysis was used to quantify skin colour, gloss, tone evenness and wrinkling/sagging. We assessed associations between face ratings and skin image measurements in the total sample (i.e. all ethnic groups) and separately by ethnicity.ResultsSkin image analysis revealed differences between ethnic groups, including skin colour, gloss, tone evenness, wrinkling and sagging. Differences in the relative predictive utility of individual skin features in accounting for ratings of age, health and attractiveness also were observed between ethnic groups. Facial wrinkling and sagging were the best predictors of face ratings in each ethnic group, with some differences in the type (or predictive magnitude) of skin features.ConclusionThe current findings corroborate previous reports of differences between ethnic groups in female facial skin and indicate differential effects of skin features on ratings of age, health and attractiveness, within and between ethnic groups. Facial wrinkling and sagging were the best predictors of age and attractiveness ratings, and skin tone evenness and gloss had an additional role in ratings of health.