From a frown to a smile, the technology that’s in your face

FT.com

January 2, 2014 4:40 pm

By Sally Davies

Face the facts: mapping by CrowdEmotion shows happiness

Shakespeare called it “a book where men may read strange matters”. What the human face reveals has fuelled speculation, incited wars and inspired poetry for centuries.

Now a clutch of companies is chipping away at the mystery of facial expression, thanks to technology that attempts to detect the feelings behind fleeting facial movements. The applications range from heartwarmingly helpful to creepily intrusive.

CrowdEmotion, a London start-up, is in talks with a baby foodmaker to license an app that scans children’s faces to discern if they are hungry or tired, for instance. Australian company Seeing Machines, which tracks the eyes of truck and bus drivers to see if they show signs of falling asleep, is also working on ways to detect when drivers are frustrated or aggressive, which could affect how they drive.

In the US, emotional analytics company Affectiva has tested advertisements for market research organisation Millward Brown, as well as raising nearly $22m from investors including WPP, the advertising group that owns Millward Brown.

Meanwhile, retailers and restaurants have expressed interest in using San Diego-based Emotient’s facial recognition software to measure customer satisfaction and employee cheerfulness. “For myriad reasons we have a difficult time expressing what we really feel,” says Ken Denman, Emotient’s chief executive. Facial expression technologies help people “have a more intuitive relationship with products, services and all things digital”.

Attempts to measure emotions have been given a boost by technological innovation. Algorithms that harness the power of cameras and cloud computing can now sift facial data frame-by-frame to gauge how people react to products and situations in the real world. This includes capturing involuntary facial flickers that may be a sign that someone is trying to hide or suppress their feelings.
The technology is most advanced in marketing, where cameras observe the expressions of volunteers as they watch videos on computers, smartphones or tablets. The more visual data these programs capture, the better they learn how tiny twitches relate to emotions.

Using the face to read feelings is a tantalising prospect for marketers. “Mood drives behaviour,” says Matthew Celuszak, CEO of CrowdEmotion. The idea is that faces offer a way to judge the effectiveness of advertising campaigns, because certain expressions correspond to consumers’ responses.

However, the link between how consumers feel and how they act is often framed simplistically, says Robert Heath, associate professor of advertising theory at the University of Bath. He warns that marketers may falsely assume the element that is measurable must be the deterministic cause of an outcome.

The rise of always-on, wearable devices could help push facial recognition mainstream, says Anthony Mullen, an analyst at technology research firm Forrester. It could catch on in gaming, for instance, paving the way for immersive worlds that respond to players’ emotions by unlocking new levels or changing pace.

As well as questions of accuracy, the technology raises privacy concerns. In response, most emotion-tracking companies abide by rules whereby people consent to have their faces read. The data can also be anonymised.

But aside from seeking permission to collect facial information, the question of how to be transparent about how the software draws inferences is an “interesting technical challenge”, says Mr Mullen. If conclusions about someone’s emotional state were used to make employment decisions, for example, it would need to be made clear to the person how the technology works and why it is valid.

The Bard’s words that “false face must hide what false heart doth know” may be true a while yet.