Captavatar – Using principles of human multisensory perception to endow virtual agents with maximum social appeal

A new collaboration between the Multisensory Cognition Group and GV2 starting September 2011 funded by Science Foundation Ireland.

There is an increasing demand for more realistic virtual characters, especially for games that involve interaction between the user and a virtual human, and entertainment applications designed to evoke intense emotions. Within the Creative Industries, there is insufficient understanding about how the perceivable aspects of virtual characters (e.g., facial expressions, voice intonations or bodily gestures and their combinations) increase user engagement. In a novel inter-disciplinary project combining computer graphics and social cognitive neuroscience, the work will apply principles of human perceptual processing to endow virtual agents with maximum social appeal. By identifying preferred features and their multisensory combinations (measured through behavioural and brain responses) the research will then simulate these features in virtual agents, and create a ‘morphable’ social human. The effectiveness of the research will be demonstrated in the areas of games and entertainment (as specified by our collaboration with Intel and Disney Research), and compelling applications for education and health-care.