A National Institute of Standards and Technology (NIST) study has found that face recognition algorithms developed after the COVID-19 pandemic’s onset have exhibited better performance in identifying masked faces than software created before the pandemic.
“Some newer algorithms from developers performed significantly better than their predecessors. In some cases, error rates decreased by as much as a factor of 10 between their pre- and post-COVID algorithms,” Mei Ngan, one of the authors of the NIST study, said in a statement published Tuesday.
The study evaluated the performance of 65 post-COVID-19 algorithms in addition to previous software tested on masked faces and employed the same set of 6.2 million images to test the ability of the algorithms to conduct “one-to-one” matching or comparison of two different pictures of the same individual.
“In the best cases, software algorithms are making errors between 2.4 and 5 percent of the time on masked faces, comparable to where the technology was in 2017 on non-masked photos,” Ngan added.
