Woads? Where we’re going, we won’t need woads.

Paul Raven @ 28-02-2011

CVdazzle make-up and haircuts for fooling facial recognition softwareLooks like this one’s resurfaced again; we mentioned Adam Harvey’s efforts to spoof facial recognition software with weird asymmetrical make-up patterns a little less than a year ago, but the project now has its own name and URL – CVdazzle.com – and has expanded into using random hairstyling techniques… which should keep the junior stylists at Toni & Guy in business once the whole emo/Bieber/Lady Gaga thing dies out.

[ Image copyright Adam Harvey; please contact for immediate takedown if required. And yes, I know the pun in the title is terrible, but a) it’s my blog and b) it’s Monday, so please address all complaints to my imaginary lawyer. 🙂 ]


Fooling facial recognition software: the Adam Ant method

Paul Raven @ 07-04-2010

This one’s been doing the rounds for a few days, but I thought I’d throw it out for those of you who’ve not seen it yet: some graduate research suggests that stylised facial decoration and make-up (a la eighties New Romo band promo shots and cyberpunk bandwagon movies) may be a simple and effective counter-response to the increasing ubiquity of surveillance cameras with facial recognition software [via Technoccult].


Good with faces? The Recognition Corps needs you!

Paul Raven @ 25-05-2009

little metal facesOK, so it’s not quite X-ray vision, but science has unearthed a slightly more prosaic super-power – super-recognisers are those rare folk with an excellent memory for faces. [image by Valeyoshino]

Super-recognizers report that they recognize other people far more often than they are recognized. For this reason, says Russell, they often compensate by pretending not to recognize someone they met in passing, so as to avoid appearing to attribute undue importance to a fleeting encounter.

“Super-recognizers have these extreme stories of recognizing people,” says Russell. “They recognize a person who was shopping in the same store with them two months ago, for example, even if they didn’t speak to the person. It doesn’t have to be a significant interaction; they really stand out in terms of their ability to remember the people who were actually less significant.

One woman in the study said she had identified another woman on the street who served as her as a waitress five years earlier in a different city. Critically, she was able to confirm that the other woman had in fact been a waitress in the different city. Often, super-recognizers are able to recognize another person despite significant changes in appearance, such as aging or a different hair color.

I’m better with faces than I am with names, though not to that extent! Worse luck for me, as Randall Parker of FuturePundit suggests an employment fast-track for super-recognisers:

Imagine employing super-recognizers to watch for criminals in train stations, airports, and other places where large numbers of people pass. These are the people we should want to look at most wanted lists of criminals.

Yeah – let’s replace those horrible cameras with fallible meat-based undercover snooper agents! After all, we can always be sure the people on the most-wanted lists deserve to be there, right?

Right?