Output details

Modelling the human blink: a computational model for use within human-robot interaction

<24> This paper shows for the first time that at least 48% of blinks are not random or physiological, but are closely correlated to communicative behaviour and changes in internal communicative states. A blink generation model is proposed that reproduces the experimental statistics and allows the animation of a face with the most natural blinking behaviour possible to date. Impact: This new results and methodology are highly relevant in the of field facial animation for human-robot interaction, and are being used for experiment with the Lightbot robot developed in Belpaeme’s EPSRC project “CONCEPT” and commercialised by the spin-off company Synthelligence.