A Computer Vision Methodology to Predict Brand Personality from Image Features
本研究用计算机视觉方法分析颜色和视觉复杂度对品牌个性的影响，提出四步流程将计算编码的视觉属性与人类对品牌个性的评分相关联，对广告研究者和从业者有方法论意义。
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.