Text this: Object representation in computer vision :