It’s making some surprising discoveries.
Photo via (cc) Flickr user via Theoklis_
Humans interpret the emotional states of others by analyzing complex, subtle gestures at a deep level. Although facial expressions have different meanings depending on cultural context and gender, new facial-recognition technology is being used to analyze these subtle signals for the detection and treatment of clinical depression.
Researchers at Carnegie Mellon are using a new system called MultiSense, whose multimodal algorithm tracks a person’s gaze, body orientation, and head position and matches those findings with 68 points on their face. Initial research with the technology has yielded some unexpected results. A surprising find was that depressed and nondepressed people smile at exactly the same frequency, but the smiles of depressed people last for shorter periods of time.
MultiSense also helped researchers discover that depressed men frown more frequently than nondepressed men, but in the case of females, depressed women actually frown less than those without depression. This type of data gives researchers new jumping-off points for delving deeper into the social and biological aspects behind facial expressions. “The really interesting next part,” says Louis-Philippe Morency, an assistant professor in Carnegie Mellon’s School of Computer Science, “is to see how [these findings are] aligned with social norms. Is it related to culture? Is it local? National? International? Or is there even another factor—social, cultural, physiological—that we don’t know yet?”
This initial research is just the tip of the iceberg for MultiSense’s potential. The Department of Defense is using it as a tool to detect PTSD and looking into ways it can be used to predict behavior. Although he’s excited by MultiSense’s future, Morency is cautious about how it might be used. “I personally would much prefer this technology as a decision-support tool, but not as a decision-making tool,” Morency said. “When we start talking about decision-making, this brings a large number of ethical issues that need to be addressed.”
(H/T The Atlantic)