Trinamix extracts three data streams at the same time: 2D image, 3D depth map and skin detection. Source: TrinamixTo thwart facial login “spoofing” attempts — where hackers attempt to gain access to facial recognition-protected smartphones and other devices using pictures, masks and even a 3D-printed head — a team from the tech company Trinamix has developed a “skin-sensing” technology to rival standard facial recognition technology.

Using what it calls “beam profile analysis” technology, Trinamix can distinguish between real, living skin and 2D images of skin and even latex and silicone masks designed to resemble human skin.

The sensor, which is intended as a future smartphone feature, is an optical module that gathers data using three components: a 1 megapixel near infrared camera, a light projector and a flood illumination sensor. Infrared beams from the module map out facial features to devise a 3D representation of a subject for verification purposes. Built-in algorithms then conduct material classification that enables the sensor to distinguish real skin from latex based on the characteristics of both.

The company developed the technology following a spate of reports where hackers were turning to creative methods for unlocking devices protected by facial recognition technology, including reports that one person had commissioned a 3D printed head in his own likeness to demonstrate that the technology could be tricked.

In addition to applications for thwarting attempts to override the facial recognition features of a device, Trinamix believes the technology could eventually be used in manufacturing, linked to robotic arms on the warehouse floor to help the robots sort items based on their material composition. For instance, a robotic arm might be capable of distinguishing between glass and plastic in a recycling facility if outfitted with the technology.

For now, the technology is still in development.