Learning to reconstruct

Suppose that you were given a blurred image: Then you could probably tell that it was generated from this image: But what about if you instead »