Building computers’ emotional intelligence
3 July 2017: Leading thinkers in human emotion, technology and computing have explored new ways for technology to monitor and interpret feelings at an event at the University of Canberra.
The research colloquium was recently hosted by the University’s Human Centred Technology Research Lab (HCT) and Professor of Affective Computing Roland Goecke.
Professor Jeffrey Cohn from the University of Pittsburgh was a key-note speaker at the event. He provided insights into his research into understanding human behavioural cues such as emotion, facial expression and body movement.
“A lot of my investigation has been around understanding some specific phenomenon often with a clinical action we want to take, like treating people with depression or autism,” Professor Cohn said.
“The approach we’ve seized on and have found useful in face analysis is a sign-based approach, which steps away from what the expression means or is interpreted as. We don’t say a person is smiling because that implies the emotion, we’d say the corners of their mouth are pulled upwards.”
Professor Cohn, an expert in psychology and psychiatry, has worked with Professor Goecke on the University’s affective computing research into depression and other mood disorders.
The work focuses on developing computer-based analysis of small physical expressions to aid mental health clinicians in diagnosing and monitoring mental health.
“Being able to gauge that emotional state via video, for example a psychologist interviewing a potential patient, provides a real measurement to back up the professional, but still intuitive opinion of the clinician,” Professor Goecke said.
Professor Goecke said the colloquium was an opportunity for University of Canberra academics and students to see some of the cutting-edge research happening around the world in interpreting human emotions through technology.
“Some of our guest speakers have been working on computer programs to read people’s faces and exploring how humans interpret emotional cues to improve technology,” Professor Goecke said.
“At HCT we are working in areas such as pattern recognition, signal processing, computational intelligence and telecommunications. Hearing about the successes our colleagues are having overseas is important as we continue to refine our work here as well.”
Other guest speakers at the colloquium included Professor Tom Gedeon from the Australian National University, Julien Epps from the University of New South Wales and Saurabh Sonkusare from QIMR Berghofer Medical Research Institute.