Their whole facial bone structure is the same, that’s the result of the training data I’m using. Great for making consistent characters in different poses and settings.