Accounting for Private Taste: Facial shape analysis of Attractiveness and Inter-individual Variance

• Published in: 2020 12th International Conference on Knowledge and Smart Technology (KST) pp. 203 - 206
• Main Authors: Tanaka, Takumi, Mikuni, Jan, Shimane, Daisuke, Nakamura, Koyo, Watanabe, Katsumi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-01-2020
• Subjects: Correlation coefficient; facial attractiveness; Facial features; Feature extraction; Forehead; individual difference; Predictive models; private taste; Robustness; sexual dimorphism
• DOI: 10.1109/KST48564.2020.9059511

Attractiveness is an important facial attribute, which could largely influence individuals' impressions or social relationship. Lately, numerous studies have examined the morphological facial features driving attractiveness. Notably, most of the studies have postulated the "ground-truth", universal standards of facial attractiveness. However, in fact, it is well-reported that there are inter-individual differences in attractiveness judgments. These individual differences may be derived from individual preferences for certain morphological features in the faces. Nevertheless, there has been no direct empirical study to investigate the variances in the evaluation of facial attractiveness. In this study, we examined the quantitative relationships between morphological facial features and the judgments of attractiveness ratings for the faces. We found that the variances in attractiveness ratings could be partly predicted by some traditional machine learning, and that the sharpness of face outlines and morphological features representing the smile expression could have impacts on the amounts of the variances.