Facing the Void: Overcoming Missing Data in Multi-View Imagery