Micro-gesture analysis with machine learning for hidden emotion understanding


Project description

Emotions are a central part of human communication, play an important role in everyday social life, and should have a key role in computer-mediated interaction. When people try to hide their true feelings, the subconscious mind acts automatically and independently of our verbal information, so body language could give additional indication. Even people who have mastered body language can make small mistakes; these mistakes are what we call micro-gestures. Micro-gestures are small, automatic gestures, most of which are beyond our awareness, or unconscious. Being able to automatically detect and then amplify such gestures often enables to discover their symbolic meaning, opening up rich paths of emotional intelligence. In this research, analysing the very fine body gesture clues with machine learning methods for detecting and recognizing suppressed or concealed emotions, can help spot and interpret them, and then people’s true feelings can be understood.

Project coordinator

University of Oulu


Guoying Zhao

Guoying Zhao