Faces and Feelings
He and his team identified 821 words that are associated with emotions, such as anger, frustration and happiness. These words and their translations were used to mine the web in popular search engines in 30 countries.
The process resulted in about 7.2 million images of faces that reflected those specific emotions. Each image was coded by a computer algorithm using a measurement called action units (AUs) which represent the muscular activity that trigger changes in facial appearance. They were sorted into categories of emotions that spanned all the cultures studied.
The human face is capable of configuring itself in 16,384 unique ways, and Martinez found 35 that were the same across all cultures. "It was reassuring to me that we are a happier society and world than we thought," Martinez says. "Almost half of the expressions we identified were positive ones."