Interactions with virtual objects for machine control

A non-transitory computer readable medium storing instructions for manipulating a virtual object, which instructions when executed by a processor perform:

• detecting a hand in images of a three-dimensional (3D) sensory space captured by at least one sensor;
• generating from the images a predictive model of the hand, wherein the predictive model is used to track motion of the hand;
• assigning in the predictive model, calculation points assigned to portions of the predictive model corresponding to positions of fingers, thumb and palm of the hand;
• dynamically selecting at least one manipulation point describing a point of interaction of the hand with virtual objects based on the motion of the hand tracked by the predictive model by:
• weighting distances between positions of one or more of the calculation points and an anchor point defined for the predictive model and selected according to a type of interaction and/or a location of where an interaction is to occur;
• determining an average distance of the weighted distances between positions of the calculation points and the anchor point;
• and
• selecting the at least one manipulation point relative to the calculation points using the average distance;
• and
• manipulating the virtual object with the hand by interaction between at least some calculation points of the predictive model and the manipulation point.

A system, including:

• one or more processors;
• at least one sensor;
• a computer readable memory coupled with the one or more processors and the at least one sensor, the computer readable memory storing instructions for manipulating a virtual object, which instructions when executed by at least one processor perform:
• detecting a hand in images of a three-dimensional (3D) sensory space captured by the at least one sensor;
• generating from the images a predictive model of the hand, wherein the predictive model is used to track motion of the hand;
• assigning in the predictive model, calculation points assigned to portions of the predictive model corresponding to positions of fingers, thumb and palm of the hand;
• dynamically selecting at least one manipulation point describing a point of interaction of the hand with virtual objects based on the motion of the hand tracked by the predictive model by:
• weighting distances between positions of one or more of the calculation points and an anchor point defined for the predictive model and selected according to a type of interaction and/or a location of where an interaction is to occur;
• determining an average distance of the weighted distances between positions of the calculation points and the anchor point;
• and
• selecting the at least one manipulation point relative to the calculation points using the average distance;
• and
• manipulating the virtual object with the hand by interaction between at least some calculation points of the predictive model and the manipulation point.