A Computer Vision Methodology to Predict Brand Personality from Image Features

Abstract

Using the computer vision method, this study proposes an analytical model of visual aesthetics for brand communication and analyzes the effects of visual features (i.e., colors and visual complexity) on brand personality. This study illustrates a four-step procedure correlating computationally coded visual attributes with human ratings of perceived brand personality. This study has important methodological implications for advertising researchers and practitioners.