Facial Recognition Technology FTR is used to identify people. This technology is used on phones. People can unlock their phones with their faces. It’s been around for a long time.
The FRT virus is controversial and people are worried that it might be dangerous. But it seems to be safer than other ways to find someone’s identity. Check out our guide on FRT, and you can make up your own mind!
WHAT IS FACIAL RECOGNITION?
In Computers, there is a thing called Computer Vision Technology. This technology finds and distinguishes faces. For these kinds of systems, an image is given a set of data with certain features. FRT algorithms rely on biometric data – facial characteristics such as hard tissue, curves of the eye socket, nose and chin.
In 2020, someone made a way for people to tell who was in a photo on social media. Now, this is a popular technology that is becoming more important for people around the world.
HOW FACIAL RECOGNITION WORKS?
1) The FRT algorithm contains two stages. The first is the identification and the second is verification.
2) The sequence of actions in this algorithm are, usually, as follows:
3) A face detection algorithm finds the person’s face in the picture.
4) A new algorithm to detect faces has been created. It calculates points on the face that define how people are different from each other. This is new because before, algorithms only looked at where people were looking.
5) To get a good picture, the FRT system does facial normalization. It does more work to get a clear picture of your face.
6) This is about a computer program that takes information and calculates things. It needs to know what the person looks like so it can check if they match with the other people. A description tells the computer what they look like. It doesn’t matter how old they are, or if they wear makeup or not – it just needs to show their face
7) Finally, verification is important. When you do verification, you compare your image with images in the database to make sure it is a match.
How accurate is facial recognition?
In December 2019, the National Institute of Standards and Technology (NIST) issued a report on FRT. They found that because of their research, facial recognition algorithms can see better results across groups and “undetectable” qualities.
This study looked into the accuracy of algorithms and the presence of bias. For example, one test result showed that black people were more likely to be misidentified than white people. Black women had a 0.1% false match rate, and black men had a 0.025% false match rate–both 10 times higher than white people
The NIST published a report in January 2020. This report said that the former assumptions about the algorithm are wrong. In fact, the most popular FRT systems do not have a significant bias.
Hence, research shows that 17 of these good FRT neural networks have shown similar recognition accuracy, regardless of race or gender
Besides, a study by NIST shows that FRT algorithms are the most accurate when they work together with humans. The date of this study is March 2017.
Researchers did research to find out how good the cameras are in good conditions. They found that they can be 99.97% accurate in ideal conditions, but they aren’t always perfect.