Not a member yet? Register for full benefits!

ACQUINE and Virtual Identity

In a handful of the more pioneering virtual environments, a system called FaceGen, along with other, similar systems, allows a user to photograph their face from front and side, and use that to put together a 3D model of their physical head, if they so desire, to use for the basis of their avatar presence online. Alternatively of course, photos of someone else might be used, or a photo might be photoshopped to have it look more attractive.

On that last point, there is now a system available called ACQUINE, which was developed at Penn State University, designed to look at such photographs, and determine their aesthetic value according to a set of standardised norms. ACQUINE or Aesthetic Quality Inference Engine - Unbiased Assessment of Photos is of course not just limited to human faces, and will assess any photograph for aesthetic nature. However, the usage of this program, and others like it, does open up an interesting facet of such face generation systems.

With ACQUINE, the system extracts and uses visual aspects such as colour saturation, colour distribution and composition to give any uploaded image a rating from zero to 100. The system learns to associate these aspects with the way humans rate photos based on thousands of previously-rated photographs in online photo-sharing Web sites. This builds up a neural net approach over time that allows the system to become more and more proficient at grading the aesthetic quality of photographic images.

A system such as this, could quite plausibly be used to judge the aesthetic quality of a face to be made into a 3D model. When combined with heuristic algorithms, offering changes slight to a face in different ways, it is easy to envisage a use where those dissatisfied with the way their face looks, will very shortly be able to morph it closer to desirable norms, using such software to tell if the 'improvements' are more, or less desirable than the original, in successive stage after successive stage.




Online system rates images by aesthetic quality

Staff Comments


Untitled Document .