Modeling and Animating Human Figures - Computer Animation: Algorithms and Techniques - page 293

fingers

1 DOF

2 DOFs

thumb

3 DOFs

palm

wrist

FIGURE 9.10
Simple model of hands and fingers.

A model similar to Figure 9.10 is used by Rijpkema and Girard [ 54 ] in their work on grasping.
Scheepers [ 57 ] uses 27 bones, but only 16 joints are movable. Others use models with subtler joints
inside the palm area in order to get human-like action.
If the hand is to be animated in detail, the designer must pay attention to types of grasp and how
the grasps are to be used. The opposable thumb provides humans with great manual dexterity: the
ability to point, grasp objects of many shapes, and exert force such as that needed to open a large
jar of pickles or a small jewelry clasp. This requires carefully designed skeletal systems. Studies
of grasping show at least 16 different categories of grasp, most involving the thumb and one
or more fingers. For a given task, the problem of choosing which grasp is best (most efficient
and/or most credible) adds much more complexity to the mere ability to form the grasp.
Approaches to grasping are either procedural (e.g., [ 5 ][ 8 ][ 30 ][ 46 ] [ 54 ]) or data driven (e.g.,
[ 18 ][ 39 ] [ 51 ]).
Simpler models combine the four fingers into one surface and may eliminate the thumb
( Figure 9.11 ) . This reduces both the display complexity and the motion control complexity. Display