See Anyone You Know? Face Recognition Comes Of Age

The National Academy of Sciences are about to publish an article in their proceedings entitled ‘Privacy In The Age Of Augmented Reality’, co-authored by Alessandro Acquisiti, Ralph Gross and Fred Stuzman. It is about developments in face recognition software.
How 2D facial scanners record identitiesTo use the authors’ words the document

“investigate(s) the feasibility of combining publicly available Web 2.0 data with off-the-shelf face recognition software for the purpose of large-scale, automated individual re-identification.”

They are also working on an app that can do it all from your phone! See the FAQ section here for more information. The article reports a series of experiments conducted over the last year or so during which the researchers try to identify a person from their photo using an over the counter face recognition software using information that is freely available over the internet.

The results are interesting. The experiments are as follows:
Students walking through the university campus were asked if their photo could be taken and to complete a questionnaire. As they were answering the questions the computation task was carried out, looking for a picture match on Facebook and requiring only seconds. In this case more than 30% of the students were immediately traced.

Because the faces were the same but the photos taken from different angles, humans had to decide which of the possible matches were the most appropriate, but that is not always the case. Some photos are replicated and therefore the computer can give a 100% guarantee that the match is correct.

For example in another experiment the researchers used an online dating agency that provided anonymous photos. In this case they could match names to the photos in about 10% of cases. In several cases the same photo had been used on different sites.

In a third experiment the knowledge gained was used to search for further private information, all freely available on the web, such as details of sexual preference, date and place of birth and this information even allowed them to generate the first five figures of the individual’s US social security number.

So it seems that we can draw a simple conclusion here, either now or in the very near future, as these technologies are improved and made freely available, anybody will be able to recognize anybody they see on the street, identify them through an app in their telephone, and find out about their interests and other personal information, if they have ever posted (or had posted for them) a photo of themselves on the internet.

For more information, please read my face recognition article on the Bassetti Foundation website.

Online Gamers as Scientists

If you thought that online gamers were just a load of geeks, incapable of socializing with the outside world, and living within the confines of their own in their bedrooms, you might like to have a look this website called Foldit. Foldit is a game, but its aim is to solve puzzles for science, and players have recently made some remarkable inroads into the world of protein modelling. Below is a model of an Amino Acid, and it is this type of thing that gamers manipulate.
An Amino Acid Protein MoleculeThis article that explains the process is in the online journal Nature, Structure and Molecular Biology, and begins with the following statement:

“Following the failure of a wide range of attempts to solve the crystal structure of M-PMV retroviral protease by molecular replacement, we challenged players of the protein folding game Foldit to produce accurate models of the protein. Remarkably, Foldit players were able to generate models of sufficient quality for successful molecular replacement and subsequent structure determination. The refined structure provides new insights for the design of antiretroviral drugs.”

The fold it game has existed for a couple of years now. Players create protein structures, with the most stable and low energy structures scoring the most points.

The gamers in general are not scientists and they manually manipulate the model from a base form that is provided to them at the start of the operation. They have a variety of tools but the most important thing is that they have better spatial reasoning skills than computers. Computer models had tried to solve the problem cited above for 10 years without success, gamers produced an adequate model that was then refined by scientists in just 3 weeks.

We could draw similarities to citizen science, having seen posts on this blog discussing loaning out some of your computer’s spare hard disk space and memory to solve scientific problems, and the now common use of similar set ups in astronomy.

Just this week the Astronomy and Telescope journal is entitled Citizen Science, and addresses the issue of amateurs classifying high definition photos of far off galaxies. They say that it is the future of astronomic discovery. See my post on The Bassetti Foundation website for a lay explanation.

The gaming process is an interesting innovation though, as it uses skills that may not be particularly associated with science, but reveal themselves to be extremely important.