Expressive virtual modalities for augmenting the perception of affective movements

Abstract

Recent progress in augmented reality enables them to be integrated in performing arts applications. So-called "augmented performances" are set up around the world, providing the audience with a new experience. As artistic tools, augmentation elicits aesthetic emotions within the audience by conveying the artists' emotional intent. The emotional communicative power of augmented reality in a performance has however not yet been formally evaluated from a perception point of view. This paper reports an experiment about the emotionally communicative power of several virtual modalities. We first describe how we collected a set of emotionally-expressive dance sequences from a professional dancer. We explain how the emotion expressed in each sequence is recognized by computer automatic recognition techniques and human subjects. We explain how we generated for each sequence virtual elements that are animated according to the expressed emotion. The resulting augmented dance sequences are then evaluated by human judges in order to determine which augmentation better conveys the original emotion. Results report that complex emotions are harder to recognize than simple emotions in a dance context, but that augmentation might improve recognition of those complex emotions.