Did you read the article? They show the image that was matched of the professor compared to the training image, they are completely different perspectives.