Animation of hand motion from target posture images using an anatomy-based hierarchical model

This paper presents an anatomy-based hierarchical approach for modelling the human hand and animating hand gesture. The resulting technique greatly simplifies the motion data acquisition and the process of motion analysis and synthesis. Key contributions of our approach include (1) no high-speed motion tracker is needed as the technique is able to generate and animate anatomically realistic intermediate motion or in-betweens from only the starting and the final static postures of hand gestures; (2) the interpolated postures generated by the synthetic hand model are anatomically realistic; and (3) a gesture database can be easily built incrementally and automatically. The use of model hierarchy abstracts the underlying representations and computations, which allows the hand motion to be specified using high-level behavioural descriptions. In this work, we also propose a hand action coding system (HACS) to codify hand gestures in terms of hand muscle activation commands. The anatomy-based model, together with the HACS, allows complex sequence of hand gestures to be animated using high-level textual scripts or from static images of the target hand gestures to be animated. As a by-product, the anatomical-based hand model can potentially be applied to inferring which muscle or group of muscles are involved in a given hand gesture.