There was a mask that deceives facial recognition systems

"We tested the effectiveness of our mask in real experiments by printing a design on it,

In these experiments, the facial recognition system was only able to identify 3.34% of the participants who wore the mask.Compared to other masks, such as medical masks, the result was 83.34%," the researchers noted.

They explained this by saying that the COVID-19 pandemicmade wearing masks a habit, which helped many facial recognition systems to adapt. Now, however, the researchers have used a gradient-based optimization process to create a mask that will help falsely classify each user, whether male or female.

Face recognition models can trainrecognize people wearing unusual masks. Alternatively, each masked face image can be pre-processed at the output stage to make it look like the person is wearing a standard mask.

The researchers relied on the fact thatface recognition compares two parts of a face and tries to answer the question of whether they belong to the same person. Therefore, scientists generated an unusual image that continues the face of the mask owner (sometimes this pattern resembles an animal). This approach confuses the algorithms, and it can only in some cases identify a person. So they generated separate drawings for each user based on their facial features.

Read more:

Scientists have found a "Pandora's box" in the bowels of the Earth: the energy from there feeds life on the planet

NASA releases first full image from James Webb Space Telescope

The main myth about dinosaurs has been refuted: scientists have understood how reptiles took over the planet