Sberbank’s website reports that the bank and its ecosystem company VisionLabs, one of the global leaders in computer vision and machine learning, have analyzed if face masks can affect security and face, recognition especially in big cities. Amid the coronavirus threat, the number of people wearing face masks increased. Will the biometric recognition system be affected, if people cover part of their faces? Does it make sense that criminals in movies always hide their faces and never use face masks?
As Alexander Khanin, Founder and Chairman of VisionLabs, noted, “Our products can recognize faces even when people put on face masks. At VisionLabs, we added this option to our product back in 2018 and implemented more than 20 projects in Southeast Asia and Japan where people wear face masks on a daily basis. Of course, face masks make the recognition process more difficult, but we’ve managed to identify people who covered their faces with fake mustaches and beards, or wear glasses, scarfs, hats, face masks or other accessories with high accuracy. Our algorithms take into account a large number of facial features to recognize a person. Only some of them will be taken into account if their face is partly hidden, but the algorithm continues to work consistently with other visible facial features.”
Rambler.ru, a Sberbank ecosystem resident, carried out a poll to learn about Russians’ attitude toward face masks on March 5–10, 2020. Only 3% of respondents said they always use masks to prevent infection; 9% wear masks sometimes or when they are ill; and 88% of participants have never used masks.
Respondents said they buy masks at offline pharmacies (20%), online pharmacies (8%), in all sorts of places (28%), and non-pharmaceutical online stores (1%). About 45% of those who use face masks change them several times a day; 18% of them use one mask throughout the whole day; 7% of respondents use one mask for several days; 30% of them are unaware that this disease prevention means should be changed constantly.