Study: Facial Recognition Is Getting Better at Identifying People in Masks
2020-12-03
LRC
TXT
大字
小字
滚动
全页
1A U.S. government study has found that facial recognition technology is getting better at identifying people wearing masks.
2The study is part of ongoing research by the U.S. Commerce Department's National Institute of Standards and Technology (NIST).
3The agency has examined the effectiveness of more than 150 facial recognition systems on people wearing face coverings.
4The systems are powered by machine learning algorithms.
5The first results of the study were announced in July, as health officials across the world urged people to wear masks to limit the spread of the coronavirus.
6New findings were released this week.
7Police agencies around the world have long used facial recognition technology to search for and help catch people accused of crimes.
8It can also be used to unlock phones or other electronic devices, and in some cases, even vehicles.
9Some robots use facial recognition technology to recognize the people they are communicating with.
10However, the wide use of masks in public has created major difficulties for such systems.
11The study looked at facial recognition systems already in use before the pandemic.
12It also looked at systems specially developed to work on masked faces.
13Developers of the technology voluntarily provide their algorithms for testing.
14The NIST said it processed a total of 6.2 million images for the experiment.
15These included pictures provided by individuals seeking U.S. immigration benefits, as well as images from border crossings of travelers entering the United States.
16People in the images were not actually wearing masks.
17So, the researchers digitally added different mask shapes to faces in the pictures for use in the study.
18In some cases, up to 70 percent of a person's face was covered in the images.
19Overall, the NIST said its research shows that the top-performing facial recognition systems fail to correctly identify unmasked individuals about 0.3 percent of the time.
20The failure rate rose to about 5 percent with masked images tested with the most effective systems.
21Many of the lower performing algorithms, however, had much higher error rates with masked images - as high as 20 to 50 percent.
22In the latest findings, researchers included results from 65 new facial recognition systems that have been developed since the start of the pandemic.
23Some of these systems performed "significantly better" than the earlier ones, the NIST's Mei Ngan said in a statement on Tuesday.
24She is a lead researcher on the project.
25"In the best cases, software algorithms are making errors between 2.4 and 5 percent of the time on masked faces," Ngan said.
26She added that this performance rate is "comparable to where the technology was in 2017 on non-masked photos."
27The researchers reported that the systems were much more effective at identifying individuals when one image of the person was masked and the other was unmasked.
28When faces were covered in both photos, failure rates rose greatly.
29Not surprisingly, the study found that round-shaped masks - which cover only the mouth and nose - led to fewer errors than wider ones that stretch across the cheeks.
30Also, masks covering the nose led to higher failure rates than those that did not.
31The new study also ran tests to see whether different colored masks would affect error rates.
32The team used red, white, black and light blue.
33The research findings suggested that generally, the red and black masks led to higher failure rates than the other colors.
34I'm Bryan Lynn.
1A U.S. government study has found that facial recognition technology is getting better at identifying people wearing masks. 2The study is part of ongoing research by the U.S. Commerce Department's National Institute of Standards and Technology (NIST). The agency has examined the effectiveness of more than 150 facial recognition systems on people wearing face coverings. The systems are powered by machine learning algorithms. 3The first results of the study were announced in July, as health officials across the world urged people to wear masks to limit the spread of the coronavirus. New findings were released this week. 4Police agencies around the world have long used facial recognition technology to search for and help catch people accused of crimes. 5It can also be used to unlock phones or other electronic devices, and in some cases, even vehicles. Some robots use facial recognition technology to recognize the people they are communicating with. 6However, the wide use of masks in public has created major difficulties for such systems. 7The study looked at facial recognition systems already in use before the pandemic. It also looked at systems specially developed to work on masked faces. Developers of the technology voluntarily provide their algorithms for testing. 8The NIST said it processed a total of 6.2 million images for the experiment. These included pictures provided by individuals seeking U.S. immigration benefits, as well as images from border crossings of travelers entering the United States. 9People in the images were not actually wearing masks. So, the researchers digitally added different mask shapes to faces in the pictures for use in the study. In some cases, up to 70 percent of a person's face was covered in the images. 10Overall, the NIST said its research shows that the top-performing facial recognition systems fail to correctly identify unmasked individuals about 0.3 percent of the time. The failure rate rose to about 5 percent with masked images tested with the most effective systems. Many of the lower performing algorithms, however, had much higher error rates with masked images - as high as 20 to 50 percent. 11In the latest findings, researchers included results from 65 new facial recognition systems that have been developed since the start of the pandemic. Some of these systems performed "significantly better" than the earlier ones, the NIST's Mei Ngan said in a statement on Tuesday. She is a lead researcher on the project. 12"In the best cases, software algorithms are making errors between 2.4 and 5 percent of the time on masked faces," Ngan said. She added that this performance rate is "comparable to where the technology was in 2017 on non-masked photos." 13The researchers reported that the systems were much more effective at identifying individuals when one image of the person was masked and the other was unmasked. When faces were covered in both photos, failure rates rose greatly. 14Not surprisingly, the study found that round-shaped masks - which cover only the mouth and nose - led to fewer errors than wider ones that stretch across the cheeks. Also, masks covering the nose led to higher failure rates than those that did not. 15The new study also ran tests to see whether different colored masks would affect error rates. The team used red, white, black and light blue. The research findings suggested that generally, the red and black masks led to higher failure rates than the other colors. 16I'm Bryan Lynn. 17Bryan Lynn wrote this story for VOA Learning English, based on reports from NIST and The Associated Press. Ashley Thompson was the editor. 18We want to hear from you. Write to us in the Comments Section, and visit our Facebook page. 19_______________________________________________________________ 20Words in This Story 21mask - n. a covering used to protect your face or cover your mouth 22algorithm - n. a set of steps that are followed in order to solve a mathematical problem or to complete a computer process 23pandemic - n. an occurrence in which a disease spreads very quickly and affects a large number of people over a wide area or throughout the world 24benefit - n. money the government gives to people who are sick, poor, unemployed, etc. 25digitally - adv. in a way that shows information in the form of an electronic image 26overall - adj. including all the people or things in a particular group or situation 27error - n. a mistake 28significant - adj. important or noticeable 29cheek - n. the soft part of the face below a person eye and between the mouth and ear