Volumetric Modeling of 3 D Human Pose from Multiple Video