NEUROFACES: UN SISTEMA INTELIGENTE PARA EL DISEÑO EMOCIONAL DE INTERLOCUTORES VIRTUALES

Abstract

Avatars are used in different contexts and situations: e-commerce, e-therapy, virtual worlds, videogames, collaborative online design... In this context, a good design of an avatar may improve the user experience. The ability of controlling the way an avatar convey messages and emotions is capital. In this work, a procedure to design avatar faces capable of conveying to the observer the most suitable sensations according to a given context is developed. The proposed system is based on a combination of genetic algorithms and artificial neural networks whose training is based on perceptual human responses to a set of faces.