Giang, ThanhMooney, RobertPeters, ChristopherSullivan, Carol O.2015-11-112015-11-1120001017-4656https://doi.org/10.2312/egs.20001018The task of animating and rendering virtual humans in real-time is challenging. One must first establish a sense of realism through appearance, and then maintain this realism through correct and plausible motion, while continually satisfying the real-time constraint imposed. One technique that aids in achieving such a task is to take advantage of the viewer’s perception to compute less accurate models when loss of accuracy is unlikely to be noticed. Traditionally, these ‘level of detail’ techniques have been used primarily for geometric modelling. However, more recently, they have also been applied to animation. This paper seeks to amalgamate animation and geometric level of detail research in order to produce an integrated framework on which to build a totally scalable system for virtual human animation.