This document summarizes a study that analyzed millions of news photos through the GDELT database and deep learning vision APIs to understand patterns in how people and topics are portrayed. Key findings include that news photos commonly feature people, who are often portrayed as neutral or smiling. The analysis found unequal gender representation and stereotyping, with women smiling more and looking younger than men. A case study on politician photos from CNN found Hillary Clinton smiling more than Bernie Sanders in some media coverage. The researchers propose deeper analysis of the relationships between photos, headlines, topics and gender representation to build tools for computational journalism.