How Does Facial Recognition Software Work?
Have you ever tried Snapchat’s facial filters, Facebook’s automatic tagging on pictures or Apple’s Face ID? Probably, most people have at least heard about those. We use Snapchat filters to entertain and communicate with friends. We upload pictures on Facebook and it automatically tags our friends. IPhone users take advantage of Apple’s Face ID to unlock their phones without having to touch any buttons. We use those features on a daily basis but do not put thought into how it all operates. All those features are examples of the facial recognition software. You may ask yourself how does that software actually function. So let me tell you more about it.
There are multiple methods in which the software works, however, there is a general framework for the operation of facial recognition. In order to identify faces, the software has a dataset of information. The software engineers feed the software data which includes the pictures and video frames where faces are present. Sometimes that data can be collected by surveillance cameras.
The software divides the picture into approximately 80 nodal points, that are simply distinguishable landmarks on the face that have some identification patterns. Common examples of the nodal points is the distance between the eyes, the shape of the cheekbones, the length of the jawline and the width of the nose. Then these nodal points create a numerical code, called a faceprint, which is added to the database. Just like having your fingerprints taken, facial recognition creates a print and is also a type of biometric software.
The software stores the data and it compares the live or digital picture to the saved data from picture or video in order to identify the person. It uses set algorithm to analyze and identify patterns and qualities on person’s face.
The facial recognition software can be used to prevent crime, to search for suspects and even to find missing people. It is also being tested in advertising and retail to improve the customer experience while shopping. The software is already being used by major airline companies like British Airways, EasyJet, Lufthansa, Delta as a mean of identification when boarding a flight. According to the article from September 18, 2018 published by Forbes, this software can be even useful for disabled people. In 2015, the mouthwash brand Listerine launched a new app which uses the facial recognition to detect a smile and let the blind people know when they are being smiled at.
This software is a major breakthrough in technology. However, it can also put people’s privacy under question. Especially when the governments use such software, people are not usually asked for consent. There is a chance that this will become a major problem in the future.
According to Batyr Nuryyev, former software engineer in an IT startup company who worked on facial recognition software and artificial intelligence infrastructure, the primary purpose of such software is security. “Face recognition software has a lot of application,” he says. “It is very powerful but is two-sided. It can help the public security and, at the same time, can ruin everyone’s privacy.”
Just recently, a suspect was identified on a camera and caught in China during the concert which gathered 60,000 people. According to the BBC article from April 13, 2018, China is a world leader in the facial recognition technology, with a huge surveillance network of 170 million CCTV cameras installed. China uses such regulations in order to remind its citizens that they cannot escape the order and the authorities. Indeed, a person can change all the identification documents; but it is much harder to change the face.
The facial recognition software, undoubtedly, collects and uses data. The General Data Protection Regulation (GDPR) in European Union prohibits the processing of the biometric data without an explicit consent. However, many countries still have no appropriate legislation which addresses the modern technological progress and privacy issues. According to the article from September 15, 2018 published by Washington Post, cameras at 15 major international airports in the US do facial scans of travelers before they leave the country. Such scans are also required for foreigners who are entering and leaving the country.
Although a major technological breakthrough, the facial recognition software is not flawless. It is still under research and development. The effectiveness of face recognition software can decrease due to factors like the position of the face, the facial expression and even the lighting. Although the software is primarily used for security, it still collects people’s data. This can be seen as a data collection without a consent that one day can end up in the hackers’ hands. As with any other software, we should always account for possible margin of error and some glitches that the system may experience.
Aila Baizhomartova is a fourth-year student in the American University in Bulgaria double majoring in Journalism and Mass Communication and Business Administration. She is interested in the development of machine learning and artificial intelligence and its use in marketing and management.
