Face masks are thwarting even the best facial recognition algorithms, study finds


NIST digitally added face masks to immigrations photos to test 89 facial recognition algorithms.


For the most up-to-date news and information about the coronavirus pandemic, visit the WHO website.

It turns out face masks aren’t just effective for preventing the spread of airborne diseases like COVID-19 — they’re also successful at blocking facial recognition algorithms, researchers found.

In a report published on Monday, the US’ National Institute of Standards and Technology, or NIST, found that face masks were thwarting even the most advanced facial recognition algorithms. Error rates varied from 5% to 50%, depending on the algorithm’s capabilities. 

Those results are troubling for the facial recognition industry, which has been scrambling to develop algorithms that can identify people through their eyes and nose alone as more and more people turn to face masks to prevent the spread of the coronavirus outbreak. 

Face masks are essential tools to limit the disease’s spread, with governments in the majority of US states requiring people to wear coverings during the pandemic. The masks have caused trouble for facial recognition software, prompting companies like Apple to push an update so Face ID can still work even when people are wearing covers.  

Facial recognition algorithms rely on getting as much data points on a person’s image as possible, and face masks tend to take away a lot of valuable information they need to identify a person. The algorithms are already sensitive enough that improper lighting or a bad angle is enough to fool the technology, and the face masks make matters worse, the study found. 

One algorithm that would have an error rate of 0.3% surged to a 5% when presented with images of people wearing masks, the study found. The study tested the effectiveness of 89 facial recognition algorithms against face masks.

The test looked at the algorithms’ “one-to-one” matching capabilities — essentially comparing one photo of a person to a different picture, but with a mask on. NIST used 6 million images for its research, and applied masks on digitally, with different variations of the coverings. 

The study also found that the more of the nose that was covered, the more likely it was to thwart the algorithms. Black masks were also more likely to fool the algorithms than blue ones, the research showed. 

NIST said this was the first of an ongoing series of tests surrounding facial recognition and face masks, and it plans to test algorithms that were created specifically for coverings later this summer. 

“With the arrival of the pandemic, we need to understand how face recognition technology deals with masked faces,” said Mei Ngan, a NIST researcher behind the report. “We have begun by focusing on how an algorithm developed before the pandemic might be affected by subjects wearing face masks.” 

Ngan said NIST expects algorithms to improve with detecting people wearing face masks in the future. 

Facial recognition researchers have been compiling photos of people wearing face masks as data for its algorithms to learn from — in some cases, without people’s knowledge. 

The NIST study used photos of people applying for immigration benefits and digitally altered mask photos of travelers entering the US, according to the report. 

Source Article