Any computer vision people that can speak to the accuracy of using a single photo for facial recognition?