Bridging the gap between human and machine vision