Vision based personality analysis using transfer learning methods

Abstract

In this study, we investigate the techniques to automatically predict the extraversion trait among the Big Five personality traits (Openness to Experience, Conscientiousness, Extraversion, Agreableness, Emotional Stability) using visual nonverbal cues in small group meetings. Our aim is to increase the prediction accuracy by transferring the knowledge that can be extracted from conversational video blogs (vlogs) in video blogging sites, where there is vast amount of data, to predict personality with visual nonverbal cues in small group settings, where obtaining data is challenging. Our experiments show that, for predicting the extraversion trait, using vlogs as part of the training data results in higher accuracies than using only the small group meeting data.