What Is EON Human?
So what is this all about? I’ve introduced this blog to some of my family and friends, and this question is always the same. What is this?! What is so cool about it? What does it mean? Having worked in the computer graphics industry for several years I can see quite a bit of potential in this new technology, EON Human. I will try to explain what it is, and where I think it could go, as far as I know it right now. This technology is so new, not a whole lot is known currently, especially since it is still not widely available to the public. Tomorrow, May 30th, there is supposed to be a big announcement and launch of this service, through which we might be able to learn much more. But until then, I’ve been able to pick up a few bits and pieces of information here and there that I’ll try to put together for you.
The first time I became aware of this technology was when I came across a YouTube video a couple years ago. I’ve posted this video as the first post on this blog. The title of the video was “3D Morphable Model Face Animation” and was posted on April 20, 2006. Since then it has received 1,248,962 views, 3539 ratings, and 1410 comments. Not bad for a YouTube video.
In sum, what the video showed was a presentation of a project in which 200 faces were 3D color scanned and combined to make what is called a “morphable face model.” This model has the combined 3D shape, color, and texture information of all the 200 faces in one. What this allowed was continuous morphing between any of the faces. It also allowed some spectacular other capabilities:
- By comparing an original face to the combined average, a caricature could be made by moving the points of the face from the average, to the original, and then increasing the distance further along those same vectors
- By creating averages of male and female faces separately, an original face could then be compared to the average male or average female, and the difference could be added or subtracted from the original creating a change in gender perception
- Using the same method, an original face can be changed along several other lines such as weight, ethnicity, age, and beauty.
- Facial expressions captured from one face can be applied to any other.
- A full facial model with 3D shape, color, and texture can be derived from a single face photograph using a completely automated algorithm to find and match aspects from the average 3D face to the face photograph.
- Once this is done, all the features of that face can be manipulated in the photo including gender, weight, expressions, age, beauty, ethnicity, etc.
- Other 3D computer graphics can be combined with the new full 3D facial model, with accurate shadows
- Illumination can be changed or adjusted on the face in the photo.
- Poses can be varied.
- A full color 3d model can be derived and estimated from a black and white photo, because of the color information stored in the morphable model.
- The technology can be applied to any type of face, even in paintings. (The Mona Lisa is shown in this video).
If some of that doesn’t make very much sense, you are probably not alone. I don’t understand some of it either. But what it does mean is that this technology can do some really cool things, and all from a single photo of a face. And the face does not even need to be facing directly forward! This means you can use just about any face to construct the 3D model.
At that time it was not known at EON Human. It was just a random video on YouTube with a random title. The only source detail we were given was at the beginning of the movie, which said:
A Morphable Model for the Synthesis of 3D Faces
Volker Blanz & Thomas Vetter
MPI for Biological Cybernetics
Tubingen, Germany
These few clues lend us some additional information. This I will talk about in the next post.
Comments
Leave a Reply
