The new $5000 EyeSee mannequins contain cameras in their eyes that are connected to facial recognition software that analyzes customers’ faces, classifying them by age, sex, race, and ethnicity. The cameras also, according to a company press release, can tell merchandisers when the store is getting crowded, and what products or clothing are attracting the most attention.

The primary goal of the seeing-eye mannequins is to retailers maximize sales by finding ways to improve product and window displays.

According to Almax, privacy is not an issue, as analyzed pictures are not saved or transmitted, so shopper data is simply aggregated without a record of specific individuals. Bloomberg says the high-tech mannequins are currently being used in three European countries as well as the U.S., and have already led store managers to revise marketing practices after learning that men who shopped early in a sale spent more than women.

Source: Almax

EyeSee mannequins in action

(That may not be a shock to most wives, since some men — OK, yours truly — are horrible shoppers when it comes to getting good deals.)

The facial recognition technology was developed by Italian company Kee Square, a spinoff from the Polytechnical University of Milan. According to its website, Kee Square develops, integrates, and customizes biometrics products for law enforcement agencies and businesses.

The question, I guess, is whether we want our dummies looking back at us, and whether we believe that any collected data will be used, managed, and disposed of responsibly. Many stores, of course, already use sophisticated video surveillance.

But there is definitely something just a little bit creepier about dummies, mannequins, and doppelgängers looking back at us.

Image credits: Almax

Currently, Fits.me only provides male robot mannequins, but a female version should be available by the end of the summer.

Fits.me is one of the 20 (mainly European) startups pitching in TheNextWeb conference startup rally in Amsterdam this week. The company started a year ago in Estonia, and the robotics technology comes from Estonia’s Tallinn University of Technology. The mannequin robots can take on up to 100,000 different body shapes, but in practice many fewer than this is needed, since human beings tend to come in a more limited range of shapes and sizes.

Fits.me works with a German company called Human Solutions, which provides a database of the most common body dimensions taken from 3D imaging of tens of thousands of people. A selection of target shapes are chosen for the robot. Not every item of clothing in every size needs to be physically tried on the robot. Instead, sizing samples are put on the robot for each of the target body shapes, and the result is photographed. On average, you need 4 garments from a brand to establish the sizing. Big brands like Gap will have 10-15 fit samples. Most clothing brands have 4-7.

When a buyer wants to see a size on his own body shape, mathematical algorithms are used to manipulate the most similar image of the robot. This is combined with information about the physical properties of the garment and how it takes on shape (some materials are more clingy or stretchy than others) to show the user a simulation of the garment on his body shape.

In a trial with a German retailer, Fits.me increased sales 3 times and reduced returns by 28%. The company just signed UK shirt maker Hawes and Curtis as its first commercial customer. Fits.me gets a percentage of revenue from sales.

The company doesn’t seem to have any direct competitors, although there are plenty of companies that provide styling or sizing advice (see figure to the left). Based in Estonia, Fits.me launched in May 2009 and has funding of 1.3 million euros via a combination of EU research grants and private funding.