Computer now recognizes 21 different human emotions
Computers are one step closer to connecting with humans on an emotional level. Researchers at Ohio State University have taught computers how to distinguish 21 different human emotions by analyzing specific facial expressions. But what really sets this research apart is the range of compound emotions included, unlike the six primary emotions studied in the past.
A compound emotion is a combination of emotions and their corresponding facial expressions, such as “happily disgusted” or “happily surprised.” To help map out these compound emotions, 130 female and 100 male participants were told different scenarios, such as smelling a bad odor, to provoke different facial expressions. The resulting 5,000 photographs were then categorized according to the Facial Action Coding System (FACS).
The FACS data was then cross-referenced to determine similarities and differences in expressions. The result was a list of 21 emotions that had the same kind of facial expressions. Researchers are hoping this cognitive analysis can be used for therapeutic treatment of certain conditions, such as PTSD or autism.