Human motion modeling and simulation by anatomical approach

Abstract

To instantly generate desired infinite realistic human motion is still a great challenge in virtual human simulation. In this paper, the novel emotion effected motion classification and anatomical motion classification are presented, as well as motion capture and parameterization methods. The framework for a novel anatomical approach to model human motion in a HTR (Hierarchical Translations and Rotations) file format is also described. This novel anatomical approach in human motion modelling has the potential to generate desired infinite human motion from a compact motion database. An architecture for the real-time generation of new motions is also proposed