Face recognition systems tested for ability to analyze masked people

Information has appeared online that specialists from the US National Institute of Standards and Technology

decided to test facial recognition systems for their ability to analyze people wearing masks.

The results of the experiment wereboth positive and negative to some extent. So, for example, in general, most of the algorithms coped well with the task. But at the same time, they still showed a much lower level of error-free recognition than they showed in the case of people who did not wear masks.

Researchers have tested 89 algorithms,which, under normal conditions, recognized faces with an accuracy of 99.7%. But as soon as the participants in the experiment put on the masks, the best algorithm showed an increase in the number of errors by 5%. Moreover, most of the algorithms showed an error rate of 20% to 50%.