Diversified virtual camera composition

The expressive use of virtual cameras and the automatic generation of cinematics within 3D environments shows potential to extend the communicative power of films into games and virtual worlds. In this paper we present a novel solution to the problem of virtual camera composition based on niching and restart evolutionary algorithms that addresses the problem of diversity in shot generation by simultaneously identifying multiple valid camera camera configurations. We asses the performance of the proposed solution against a set of state-of-the-art algorithms in virtual camera optimisation.