Scientists Develop New Method to Fool Facial Recognition Systems

Facial recognition systems are becoming increasingly sophisticated and are now widely used in everyday life. For example, last year Apple released the iPhone X, which features the biometric Face ID system. However, these systems can still be tricked, particularly with the use of infrared LEDs. While infrared light is invisible to the human eye, most cameras can detect infrared signals.

Chinese researchers have developed a baseball cap equipped with miniature infrared LEDs. These LEDs are positioned in such a way that the infrared beams projected onto the wearer’s face not only help conceal their identity but can also allow them to impersonate another person for facial recognition-based authentication. This task is more complex and requires the use of a deep neural network to recognize a static image of a face and accurately project the infrared beams onto the imposter’s face.

To test their theory, the researchers used photos of four random individuals. They managed to fool facial recognition systems in 70% of cases, provided there was a slight resemblance between the victim and the imposter.

“Based on our findings and attacks, we can conclude that current facial recognition technologies cannot be considered safe and reliable for critical scenarios such as authentication and surveillance,” the researchers stated. They also noted that infrared LEDs can be hidden not only in baseball caps, but also in umbrellas, hair, or wigs.

It’s worth noting that shortly after the release of the iPhone X, Vietnamese researchers managed to unlock the Face ID system on their first attempt using a specially crafted mask.