Text this: Attention and Performance in Computational Vision