Cal Lightman, the lie-detecting psychologist from the Lie to me television series, may soon lose his job. Indeed, researchers at Ohio State University in the United States have succeeded in making computers identify 21 different facial expressions, three times more than until now. According to the researchers, this advance in the field of cognitive analysis could make it possible to facilitate the diagnosis of diseases such asautism or post-traumatic stress disorder.
“It is simply astounding. It means that these 21 emotions are expressed in the same way by almost everyone, at least in our culture”, remarks Aleix Martinez, researcher in cognitive sciences and principal author of these works, quoted by theAFP. In their study, published by the specialized journal PNAS, scientists say they have expanded the number of expressions recognizable by computers through the six basic emotions: joy, sadness, fear, anger, surprise and disgust.
The expressions of 230 volunteers were then photographed and analysis of a total of 5,000 images allowed computers to detect variations in major facial muscles such as the corners of the lips and the outer part of the eyebrows. By combining these data with a tool specialized in the analysis of body language, the Facial action coding system (FACS), 21 emotions composed of the six basic and 15 combinations, were identified.
