How does a new AI image recognition system determine how people are feeling?

Answer: through the analysis of micro-expressions

by News Staff / November 17, 2015

People lie about how they actually feel every day, but scientists at the University of Oulu in Finland have developed a new system that can read poker faces better than humans can. The technology works by drawing from an extensive database of labeled facial expressions (FE) and micro-expressions (ME).

"An ME is a very brief, involuntary FE which shows the emotion that people try to conceal," the authors wrote. "In contrast to ordinary FEs, MEs are very short (1/3 to 1/25 second), the precise length definition varies, and the intensities of involved muscle movements are subtle."

The technology could assist in law enforcement's use of lie detection technology, with some psychotherapy applications — or to cheat at poker.