Apple Blog

circumvent facial recognition of airports and shops in a test

Since the emergence of security systems based on facial recognition, different researchers have tried to circumvent them. Recently, a team from the Kneron startup, has carried out several tests in airports and shops equipped with this technology. The company has managed to circumvent them using a 3D printed face mask or a simple photograph, but has been unable to get past the security of Face ID .

Airports, trains and shops with facial recognition scored

Kneron is a startup based in San Diego , California. Founded in 2015, it has investments from Alibaba, Qualcomm and Sequoia Capital, among others. Interestingly, San Diego is also home to Qualcomm’s headquarters. Its mission is to create software and hardware in what they call “borderline artificial intelligence” for application in different sectors such as security, surveillance, home automation and robotics.

circumvent facial recognition of airports and shops in a test
circumvent facial recognition of airports and shops in a test

In researching their own facial recognition technology, they have tested existing ones. According to reports from Fortune, the startup has managed to overcome the security of various means of facial recognition through masks printed in 3D and even simple photographs displayed from a mobile phone . Specifically, the Kneron team used a high-quality 3D printed mask to enable the AliPay and WeChat systems to accept payments in stores in Asia.

In both cases, the subject’s face is required to be entered into the system in order to circumvent it

These systems only require the buyer’s face to execute the payment of a purchase in a physical store. They do not use PINs or fingerprints as a security measure. In addition, they have also managed to penetrate the terminal at Schiphol airport in the Netherlands, as well as various train stations in China. However, in both cases it was enough to show a photograph of the person from a mobile phone screen.

With these tests, Kneron highlights the lack of security of certain facial recognition systems. In the first case, it requires a 3D mask created from a model, something much more sophisticated than showing a photo to the scanner. Although the gravity is not the same. It is much more worrying to overcome the security of an airport or means of transport (something that can be taken advantage of by terrorists and criminals) than to pay in a shop with someone else’s bill.

Face ID maintains type in tests

In the tests, the company has used other facial recognition systems. Among them they mention Face ID only to say that they have not managed to circumvent this system . From what is mentioned, it seems that they have used at least two methods to try to overcome the different recognition systems: the photograph and the 3D mask.

In the case of Face ID, its sensors allow it to analyze the 3D relief of the user’s face, so a simple photo would not work. However, in the past we have seen tests where a mask was created that outperformed Face ID. However, it required numerous images of the subject to build a model made with 2D images and silicon on a 3D printed mask.

It is also worth remembering that test that was done at the end of last year, where the facial systems of five smartphones were analysed. The only handset that passed the test was the iPhone X . This is why Android manufacturers do not usually recommend this method of authentication for mobile payments or banking apps.

In the end, Apple made the right decision by choosing Face ID over on-screen fingerprint sensors

It is striking how Face ID still maintains the type in these tests against other facial recognition systems. Although the technology is advancing and becoming cheaper, the security levels have not increased or become more demanding.

Similar Posts