Human Multi-View Synthesis from a Single-View Model:Transferred Body and Face Representations