Metasurface-based Alternative to 3D Face Mapping Is Much More Efficient: Researchers

Researchers in Taiwan have developed a new 3D face imaging system that uses advanced, compact optical components to create a more efficient and smaller facial recognition system compared to the conventional system used in iPhones and other contemporary smartphones.

Traditional systems rely on structured light technology, which utilizes a complex assembly of a laser source, collimators, light guides, and diffractive optical elements (DOEs) to project a pseudorandom pattern of infrared dots onto the subject’s face. (Collimators are optical devices designed to narrow a beam of particles or waves, leading to parallel rays and reducing the spread of the beam.) The pattern, once reflected back, is analyzed by the device’s sensors to authenticate the user’s identity based on the unique facial contours.

In contrast, the novel method introduced by the researchers simplifies this process by leveraging metasurfaces combined with photonic crystal surface-emitting lasers (PCSELs).

Metasurfaces are ultra-thin, engineered materials that manipulate light at the nanoscale, altering its phase, amplitude, and polarization with precision. Comprising an array of nanostructures, these surfaces can shape and direct light in ways traditional optics cannot, enabling a wide array of applications from advanced imaging systems to compact sensors.

Photonic crystals are materials designed with a regular pattern, much like a microscopic lattice, that affects how light moves through or reflects off them. Think of them as a maze for light, where the structure can block some light paths while allowing others through, based on the pattern’s size and shape relative to the light’s wavelength. This ability to control light makes photonic crystals very useful for manipulating light in precise ways.

Photonic Crystal Surface-Emitting Lasers (PCSELs) are advanced lasers that leverage the unique properties of photonic crystals to direct light emission perpendicular to the crystal’s surface, unlike traditional edge-emitting lasers. This allows for highly collimated and efficient light output from a compact device.

Leveraging metasurfaces together with PCSELs allows for a much denser pattern of infrared dots to be projected from a device significantly smaller and more energy-efficient than those based on traditional technology. The result is a compact, lens-free system capable of detailed depth sensing and facial recognition, which not only reduces the physical footprint and power consumption of the device but also potentially enhances the accuracy and reliability of the recognition process.

The researchers, affiliated with the National Yang Ming Chiao Tung University and the Hon Hai Research Institute in Taiwan, say that they tested their system by scanning a replica bust of Michelangelo’s David, and that it works as well as conventional smartphone facial recognition. In fact, they say that their system generated over 45,000 dots from a device that is 233 times smaller than the surface area of a standard dot projector.

Sources: Ars Technica, Nano Letters

–

February 20, 2024 – by the FindBiometrics Editorial Team