Scientists from National Yang Ming Chiaotong University in Taiwan tested a new 3D facial recognition system and identified Michelangelo’s David, one of the most famous sculptures of the Renaissance. The study was published in the scientific journal magazine Nano Letters.
It is used in three-dimensional surface scanning, computer vision and autonomous driving systems. Such devices typically use a point projector containing several components: a laser, a light guide, and a diffractive optical element (DOE).
DOE is a special type of lens that splits the laser beam into an array of approximately 32 thousand infrared points. When a person looks at the camera, the system places these dots over a large part of their face and the camera reads a pattern to verify the person’s identity. But dot projectors are relatively large for smartphones and other portable devices.
Taiwanese experts solved this problem by replacing the traditional dot projector with a low-power laser and thin gallium arsenide plate; This significantly reduced the size and energy consumption of the structure. They applied nanostructured pillars to the top of the plate and scattered the light as it passed through the metal.
The prototype split the laser beam into 45.7 thousand infrared points. Like a traditional dot projector, the new system includes a camera to read the pattern.
During tests, the scattered laser accurately identified a three-dimensional replica of the head of Michelangelo’s David by comparing the position of dots on the sculpture with photos of the masterpiece taken from the Internet.
According to the authors of the development, their system consumes 5-10 times less energy and takes up 230 times less space than a traditional dot projector.
Previously in Russia developed Neural network to uncover deepfakes.