Towards the Automation of Binocular Depth Perception