I’d go a step further, what would be cool if you could import few pictures of a person, so one from the front, then from the sides etc. and then MetaHuman would try and make the shape of the head and face correct
there are a few techpapers showing that people are working on such tech (the 2 minute papers youtube channel shows them off every now and then). but that does not automatically mean that such reaches unreal. such things can be patented, licensed or kept for the inventors themselves.