An exclusive gaming industry community targeted
to, and designed for Professionals, Businesses
and Students in the sectors and industries
of Gaming, New Media and the Web, all closely
related with it's Business and Industry.
A Rich content driven service including articles,
contributed discussion, news, reviews, networking, downloads,
and debate.
We strive to cater for cultural influencers,
technology decision makers, early adopters and business leaders in the gaming industry.
A medium to share your or contribute your ideas,
experiences, questions and point of view or network
with other colleagues here at iVirtua Community.
According to Haaretz, an Israeli team of computer scientists has developed a software that ranks facial attractivenessof women. Instead of identifying basic facial characteristics, thissoftware has been designed to make aesthetic judgments — aftertraining. The lead researcher said this program ‘constitutes asubstantial advance in the development of artificial intelligence.’ Itis interesting to note that the researchers focused on women only.Apparently, men’ faces are more difficult to grade. But read more…
The picture on the left shows how the system is initiallycalibrated: “Facial coordinates with hair and skin sample regions asrepresented by the facial feature extractor. Coordinates are used forcalculating geometric features and asymmetry. Sample regions are usedfor extracting color values and smoothness.” (Credit: Amit Kagian, TelAviv University, Israel).
This software has been developed by Amit Kagian, a Tel AvivUniversity (TAU) student, for his master’s thesis in computer science.He has been supervised by Gideon Dror, an associate professor in computer science at the Academic College of Tel-Aviv-Yaffo and Eytan Ruppin, a TAU professor who manages the Complex Network Systems Lab.
Here are some details about how the software was tested. “In thefirst stage, 30 human participants were asked to rate from 1-7 thebeauty of several dozen pictures. Participants did not say why theyranked certain faces as more beautiful than others. The pictures werethen processed and mathematically mapped. ‘We came up with 98 numbersthat represent the geometric shape of the face, as well ascharacteristics like hair color, smoothness of skin and facialsymmetry,’ Kagian explains. Participants’ rankings of the pictures werealso input in the computer.”
But what was the second stage? “‘We input new pictures of faces intothe computer and it graded them based on the information it had.’ Humansubjects were then asked to rank the new pictures too. ‘The computerproduced impressive results: the rankings werevery similar to the rankings people gave.’ According to Kagian, the keyachievement is that the computer operated according to certainperceptions of beauty that were not input into it, but learned byprocessing the data it received.”
For more information, the researchers published their latest results in Vision Research,an Elsevier journal, under the name “A machine learning predictor offacial attractiveness revealing human-like psychophysical biases”(Volume 48, Issue 2, January 2008, Pages 235-243).
Here is a link to the abstract.“Recent psychological studies have strongly suggested that humans sharecommon visual preferences for facial attractiveness. Here, we present alearning model that automatically extracts measurements of facialfeatures from raw images and obtains human-level performance inpredicting facial attractiveness ratings. The machine’s ratings arehighly correlated with mean human ratings, markedly improving on recentmachine learning studies of this task. Simulated psychophysicalexperiments with virtually manipulated images reveal preferences in themachine’s judgments that are remarkably similar to those of humans.”And here is a link to the full paper (PDF format, 10 pages, 625 KB).
And here is a paragraph excerpted from the conclusions. “Ouranalysis has revealed that symmetry is strongly related to theattractiveness of averaged faces, but is definitely not the only factorin the equation since about half of the image-features relate to theratings of averaged composites in a similar manner as the symmetrymeasure. This suggests that a general movement of features towardattractiveness, rather than a simple increase in symmetry, isresponsible for the attractiveness of averaged faces.”
The same researchers presented their previous results at the NeuralInformation Processing Systems (NIPS) conference held in Vancouver,Canada, on December 4-9, 2006. Here is a link to
this presentationcalled “A Humanlike Predictor of Facial Attractiveness” (PDF format, 8pages, 78 KB). Here is the first paragraph. “This work presents amethod for estimating human facial attractiveness, based on supervisedlearning techniques. Numerous facial features that describe facialgeometry, color and texture, combined with an average humanattractiveness score for each facial image, are used to train variouspredictors. Facial attractiveness ratings produced by the finalpredictor are found to be highly correlated with human ratings,markedly improving previous machine learning achievements. Simulatedpsychophysical experiments with virtually manipulated images revealpreferences in the machine’s judgments which are remarkably similar tothose of humans.”
As you can see, there some shared words between these two works. The figure above is featured in both papers.
Finally, why did the researchers limit themselves to women? Haaretz says men’s faces are more difficult to rank.
Sources: Ofri Ilani, Haaretz, Israel, March 21, 2008; and various websites You’ll find related stories by following the links below.