Articulated hand configuration and rotation estimation using extended torus manifold embedding

Due to the high degree of freedom found in hand motion, it is difficult to model articulated hand configurations. In addition, observed hand shapes vary according to the hand rotation, even when using the same hand configuration. This paper presents a new manifold embedding method for modeling low dimensional hand configurations and hand rotation using a 4D torus manifold, in which the product of three circular manifolds are used. Hand shapes are extracted from 3D depth and skin color models. The experiment results using synthetic motion and real hand motions from the Kinect sensor show simultaneous estimations of hand configuration and hand rotation parameters using particle filters on the extended torus manifold.