Wednesday, April 2, 2008

Software for Facial Attractiveness

Recently there have been many discussions on a software an Israeli team of computer scientists has developed that ranks facial attractiveness of women first reported in Haaretz, some links include: Haaretz; Geekpicks; Fuzzlinks; CautionInc; ZDNet; Radio Vice; Tech News Daily; Chat Marchet.

One of the most common questions asked by reporters and bloggers is why the software only focuses on female faces? There are, in fact, several reasons for why the researchers made the software focus on women only. The research and software build directly on my PhD work found in "Is an Objective Measuring System for Facial Attractiveness Possible?" available at Amazon.

In that work it is shown that objective models of attractiveness are far more accurate for female faces than for male faces. Software for male faces is possible but would be very unlikely to be as accurate. One of several underlying reasons for this, for example, is that male faces vary in attractiveness to female observers depending on the point they are at in their menstrual cycle. At mid-cycle or ovulation, women find more masculine faces more attractive and at menses, they prefer slightly more feminine male faces. This also correlates with when women are more likely to cheat on current partners.

No comments: