This is a little touchy because I don’t think people are comfortable with the word “average” when it comes to beauty. The results of the images have some upset because the “average” face of women from around the world are actually all quite beautiful. Some think maybe too beautiful to be accurate. Nevertheless, it’s an interesting project by FaceResearch.org. Apparently, there’s no information on how the models were selected or from what I can see even how many were used. It’d be interesting to see one done where more of the statical data and original images were available.
You can check out the original article here . 
And yet they all have perfectly manicured eyebrows. Whooda thunk it!?! 😉
If you look at their website, faceResearch.org there is a demo that you can click on different faces and merge them. I did 49 different women of different races and it came out as a very beautiful woman. The hair was a little funky but the face was great.
This is amazing! I tried the demo, very cool.
It does not convince me. If you analize the faces, there is mostly a frame that serves as an ethnic pattern with zoom variations and adds on to illustrate nationalities. Follow them not in a sequence, do diagonals and you will see what I mean.
This is not surprising at all. Researchers have long known that there is a standard of beauty that crosses all cultural boundaries – that of symmetry – the more symmetrical the face the more it’s consider beautiful. So, the obvious mathematical consequence is that if you average a whole bunch of faces, you get a lot more symmetry as the grand total average. This concept is taught in most university intro psychology courses. Now with digital imaging techniques, they were able to really test this finding a lot easier. In "our days" they had to use slide transparencies and overlay multiple faces to show us examples. Here’s an article from 2006 from Boston University with less number of faces averaged: http://www.bu.edu/phpbin/news-cms/news/?dept=1127&id=41272