View-Invariant Representation and Recognition of Actions