Algorithm can read a person's face

Posted by Kate Taylor

Researchers have developed software that can tell a person's personality and mood from their facial expression.

Human beings assess one another all the time on the basis of expression, says Mario Rojas of the Autonomous University of Barcelona.

"The perception of dominance has been shown to be an important part of social roles at different stages of life, and to play a role in mate selection," he says.

The team developed an algorithm to try and predict facial trait judgments  such as attractive, competent, trustworthy, dominant, mean, frightening, extroverted, threatening, and likable. They used machine learning techniques, giving the computer a series of examples to learn from.

They trained and tested their algorithm on a set of synthetic facial images generated in a previous study by asking people to describe and rate a set of facial images associated with specific traits such as trustworthiness or dominance.

In the current study, the researchers used a subset of these images, together with their labels, to 'teach' the computer how to read a face, and tested the prediction accuracy using the rest of the images.
Three traits, dominant, threatening and mean, were found to be predictable with accuracies between 91 percent and 96 percent.

They also tested their program's predictive ability against the faces of a number of celebrities - and found that their results were highly consistent with our general ideas about these public figures.

The study also attempted to establish what information is generally used to try and predict a person's mood or character. They found that the area around the eyes contains most information about attractiveness, while the area around the mouth gives more away about extroversion.