Research project Production and perception of multimodal emotion expressions using machine learning methods

When people interact with each other, they often convey emotions through their facial expressions, bodily gestures, and tone of voice. Knowledge about nonverbal communication of emotion is important for social interactions, and has applications in many fields ranging from psychotherapy to human-computer interaction. Research on emotion expressions also provides an important source of data for theories about the nature of emotions.

Expression of emotion is investigated by measuring nonverbal “cues” (e.g., facial gestures, gaze patterns, body posture, head movements, and speech prosody), and by using machine learning methods to detect multimodal expressive cue patterns. We also conduct behavioral and brain imaging studies to investigate individual differences in emotion recognition ability.