replace face in the first start image with the face from the second start image, match face size and position of the first image, reduce head size to match head size and position of the first image, keep facial expression and hair style of the second image