A sleeker facial-recognition technology is tested on Michelangelo’s David

Advertisement: Click here to learn how to Generate Art From Text

Facial recognition systems are well-known to many people. They unlock smartphones, game consoles and online banking accounts. But the current technology is limited by boxy lenses and projectors. Researchers report their findings in ACS. Nano LettersA sleeker, 3D surface imaging device with simpler optics and a flatter surface. In proof of concept demonstrations the new system recognized Michelangelo’s David as well as existing smartphone systems.

3D surface images are used in many applications, including facial recognition in smartphones, computer vision and autonomous vehicles. These systems consist of a projector with multiple components including a lens, a lightguide, and a diffractive element (DOE). The DOE, a special type of lens, breaks the laser into an array of approximately 32,000 infrared spots. When a user looks at a locked-screen, the facial recognition software projects a series of dots across their face. The device camera then reads the pattern to confirm identity. Dot projectors are too large for small devices like smartphones. Yu-Heng Hong and colleagues, HaoChung Kuo and Yao-Wei Huang, set out to create a facial recognition system which would be compact, flat, and use less energy.

Researchers replaced a dot projector by a low-power, flat gallium arsenide plate. This allowed them to reduce the size of the imaging device and its power consumption. They etched a nanopillar design on the top surface of this thin metallic material, which creates an metasurface that scatters the light as it passes. In this prototype the low-powered light is scattered into 45.700 infrared spots that are projected onto a surface or object positioned in front. The new system is similar to the dot projector, but it also incorporates a camera that can read the patterns created by the infrared spots.

In tests on the prototype, it was able identify a 3D copy of Michelangelo’s David accurately by comparing online photos. This was achieved with five to ten times less power, and on a smaller platform (about 230x smaller) than a typical dot projector system. The researchers claim that their prototype demonstrates how metasurfaces can be used to create low-power, small-scale imaging solutions for facial identification, robotics, and extended reality.

The authors acknowledge funding provided by Hon Hai Precision Industry and the National Science and Technology Council of Taiwan. They also acknowledge the Ministry of Education of Taiwan.

Leave a Reply

Your email address will not be published. Required fields are marked *