Tech Explained: Facial Recognition

How does facial recognition work, and how can we protect our privacy?

Facial recognition, which once seemed like the stuff of science fiction, is now a part of everyday life. The technology is used in casinos, airports, mobile phones and even schools. So, how does facial recognition work, and how can we protect our privacy?

Early types of facial recognition

One of the first systems for facial recognition was developed by Woodrow Wilson Bledsoe.  Bledsoe’s system for classifying photos involved a device to input coordinates on a grid with a stylus that emitted electromagnetic pulses. The system manually recorded the coordinates of facial features such as the nose and mouth. When the system was given a photograph of an individual, it could retrieve the images from the database that most closely resembled it.

In the 1990s, the U.S. Defense Advanced Research Projects Agency (DARPA) and the National Institute of Standards and Technology designed a face recognition program which eventually led to the development of automatic face recognition technology.

One early version was tested at the 2002 Super Bowl, when law enforcement officials scanned the crowd and found several “petty criminals”. However, overall, the test was seen as a failure due to the high number of false positives.

Around 2009, law enforcement officers in the U.S. began to take pictures of suspects that could be cross-checked against a database of drivers’ licences. The following year, Facebook began using facial recognition to identify people in the photos uploaded to the site. In 2011, face recognition was installed in Panama’s Tocumen airport to try and identify drug smugglers and led to the apprehension of several suspects wanted by Interpol. By 2014, face recognition systems were being regularly used by police to identify suspects in the field.

Face recognition went mainstream for consumers in 2017, when Apple released the iPhone X, advertising face recognition as one of its primary security features. The iPhone X sold out almost instantly. Today, several companies are developing WatchList as a Service, using face recognition data platforms to help prevent shoplifting and violent crime. The global facial recognition market totaled $3.8 billion (around €3.6 billion) in 2020 and it is forecast to hit $8.5 billion (around €8 billion) by 2025.

How facial recognition works

The technology for facial recognition has improved considerably since the 1960s, but the process is still similar. Today’s facial recognition software uses either a two-dimensional image (a photo or video image) or captures a three-dimensional image from a live person. The software then measures a variety of facial features, called landmarks or nodal points on the face. These could include the distance between the eyes, the width of the nose, depth of eye sockets, distance from forehead to chin, etc. Each program uses different nodal points, and may collect up to 80 different measurements. This information is then converted into a mathematical formula which represents your unique facial signature.

To identify someone, the facial signature is then compared to a database of known faces. According to a study by Georgetown University, around half of all American adults have their images stored in a facial recognition database accessible to law enforcement.

Other uses of face recognition

In addition to apprehending criminals and securing airports and public places, there are a variety of other potential uses for this technology. Facial recognition software is being used to take attendance in classrooms and during exams. In Vienna, some crosswalk lights use facial recognition to anticipate accidents. An American startup is using the technology for machines dispensing legal cannabis. Some companies have also traded in security badges for facial recognition systems. Farmers have even used the technology to track the health of fish.

One of the largest growth areas is in retail, where face recognition can not only help stop thieves, but can also identify customers in need of assistance. Scanning the audiences at concerts and other public gatherings, could allow marketers to target products to specific groups by gender or age, for example. Banks and ATMs could use face recognition to reduce fraud and make deposits and withdrawals easier. Even restaurants could use face recognition, to serve customers or automate ordering and delivery.

What facial recognition means for privacy

The biggest issue with facial recognition is privacy – growing numbers of people are beginning to realise that it may not be in their interest to always be identifiable. Data storage systems are also under threat from hackers intent on ID theft. It has also become almost impossible to find out who has access to your image, or even the right to use your image. Issues with mistaken identity have already cropped up, and many people do not like the idea of the government (and others) tracking their every movement 24/7. 

To combat this, researchers have already developed anti-facial recognition glasses to make wearers undetectable (although security company Symantec has funded research to prevent such evasive manoeuvres). For now, concerned individuals can opt out of facial recognition systems on social media. They may also want to consider the so-called Internet of Things — those devices at home that connect to the internet. IoT devices that use face recognition include iPads, Xboxes, and smart remotes.