A Computational Personality Traits Analysis Based on Facial Geometric Features

This paper investigates the association between personality traits and facial attributes. A facial image-based computational personality traits classification is conducted to verify the connection between facial attributes and personality traits. Firstly, a database consisting of the facial images and corresponding personality traits score are collected, and facial attributes are defined based on automatic detected facial landmarks. Secondly, the classifiers are trained based on the extracted facial attributes to perform personality traits classification. The proposed method is then conducted in one-factor and two-factor personality traits classification. The one-factor classification results shows that personality traits conscientiousness and neuroticism are significantly associated with extracted facial attributes. Furthermore, the two-factor classification results reveal that the personality traits with higher heritability have more concrete association with internal facial features.