Automated Visual Coding of Politicians in the Mass Media: Correlating Large-Scale Image Analysis With Public Opinion Polls

• In Event: The Visual in Politics: Rhetorics, Frames, Effects

Abstract

Modern mass media make extensive use of images to persuade people to make commercial and political decisions. These effects and techniques are widely studied, but behavioral studies do not scale to massive datasets. In this paper, we propose introducing techniques from computer vision to facilitate automated human attribute recognition in mass media photographs. We extend the significant advances in syntactic analysis in computer vision to the higher-level challenge of understanding the underlying communicative intent implied in images. We begin by identifying nine dimensions of persuasive intent latent in images of politicians, such as "energetic" and "trustworthy," and propose a hierarchical model that builds on the layer of syntactical attributes to predict the intents presented in the images. We show that the resulting favorability judgments correlate in informative ways with independent measures of public opinion, proposing the contrasting media behaviors of visual agenda-setting and mirroring. This study demonstrates that a systematic focus on visual persuasion opens up the field of computer vision to a new class of investigations around mediated images, intersecting with media analysis, psychology, and political communication.

Authors

• Jungseock Joo, University of California, Los Angeles

• Francis F. Steen, U of California - Los Angeles