Vision in Action: Learning Active Perception from Human Demonstrations