Monday was a happier day than Sunday — at least for the subjects shown in The Guardian’s daily photo feature, “24 hours in pictures”.
A new website called The Emotional Breakdown looks at this page on The Guardian’s website every day and categorizes the facial expressions in its photos as happy, sad, angry, surprised or neutral. In Monday’s slideshow, for instance, 25% of the expressions were happy. That’s an improvement over Sunday, when 40% of the site’s expressions were angry.
The tool also works when you type in the URL of other sites. A Boston Globe photo blog about the London Riots contains more expressions categorized as sadness rather than the anger one might expect.
Plugging in fashion blog The Sartorialist returns a graph dominated by neutral and angry expressions.
“I guess you can’t be high fashion if you’re smiling or surprised,” says Andy Mangold, one of The Emotional Breakdown’s programmers.
Mangold and his partner Anthony Mattox created the site at Photo Hack Day in New York last weekend. They used Google’s Charts API to create the graph and Face.com’s facial recognition API in order to analyze the emotions in the photographs.
“It was really interesting how [Face.com] built a way to quantify facial expressions, which is something very qualitative,” Mangold says.
Many of the 40 other teams at the hack day also found interesting ways to use the Face.com API. Facialytics, for instance, uses the API to graph facial expressions in movie clips.
After abandoning an idea to track the stock market based on broker expressions, Mangold and Mattox settled on the idea to use the Face.com API with The Guardian‘s daily photo summary.
“We wanted to do something that was relevant to what was going on in the world at that time,” Mangold says.
Image courtesy of iStockphoto, mooneydriver
0 comments:
Post a Comment