Facial recognition tools fooled by Facebook photos

Researchers create 3D versions from social media pics that hack systems

North Carolina: While facial recognition technology has mostly been seen as a boon, researchers have now been able to show how such technology — using 3D rendering and light Internet stalking — can be used to steal your face, per se. Recently, at the Usenix security conference, a team of security and vision specialists from the University of North Carolina presented a system that used digital 3D facial models based on publicly available photos, such as on Facebook, and displayed them with mobile virtual reality technology, that was able to defeat four of the five facial recognition systems.

The study proved as a reminder that most bodily features remain constant, meaning that if your biometric data is compromised or publicly available, it’s can be exploited. And such information is now almost readily available for a huge chunk of the general public, thanks to social media sites such as Facebook, Twitter and Instagram.

While other groups have done similar research, the UNC tests were done collecting images only through online portals instead of physically taking photographs of volunteers, with the group saying they were able to find anywhere between three and 27 photos of volunteers. True Price, one of the researchers, said that the group then used the photos to identify “landmarks” on each person’s face, fit these to a 3D render, and then used the best quality photo to combine data about the texture of the face with the 3D shape.

( Source : Agencies )
Next Story