Low dimensionality space for controlling human hand models

The main objective of this paper is to analyze in some depth the kinematic behaviour of the human hand, in order to obtain a low dimensionality space from the degrees of freedom most important involved in the human grasping behaviour. This low dimensionality allows reconstructing gestures with 24 degrees of freedom. The principal degrees of freedom are obtained by means of Principal Component Analysis (PCA). This analysis is carried out using a human hand model with 24 DoF, and a sensorized glove (Cyberglove). Power and precision grasps are rendered for the grasping analysis. Finally, simplified hand models are reconstructed using kinematic constraints.