Face recognition: how it works and what will happen next?

You climb the stairs and enter the elevator. He knows which floor you need. The doors to the apartment themselves open before you. The computer and telephone "recognize" you and do not require a password. Cars, social networks, shops - everyone greets you, barely seeing them, calls you by name and predicts your every move. This is how facial recognition works. Like? Scary?

At first glance it may seem that anyan organization that can afford it, monitors your every move, collects a dossier on you. But you can’t even imagine how widely facial recognition technologies have spread throughout the world and what powerful promises they promise. In addition to the above examples, face recognition systems allow you to do such simple and complex things:

  • student identity verification during online exams;
  • definition of people from the "black list" at the entrance to the stadiums and nightclubs;
  • payment for goods;
  • keeping your place in line when visiting an amusement park;
  • unlock phone or computer.

What can I say, if in Moscow alonethere is a network of more than 150,000 outdoor surveillance cameras. There is no hiding from them, and this makes people think, but the scale of “surveillance” is not so great. The network uses a powerful face recognition system, but it needs a lot of energy to work, so only 2-4 thousand cameras work in real time. Mass monitoring of the population is only scary for now, so you should focus on the real advantages of this technology. But first things first.

How does face recognition work?

Never thought about how you yourselfrecognize face, recognize it? And how does a computer do it? Of course, human faces have certain properties that are easy to describe. The distance between the eyes, the position and width of the nose, the shape of the eyebrows and chin - all these details you notice unconsciously when you look at another person. The computer does all this with a certain efficiency and accuracy, because, combining all these metrics, it gets the mathematical formula of a human face.

So, how well does the system work?face recognition now? Quite good, but sometimes wrong. If you have ever come across software that recognizes faces on Facebook or on another platform, you probably noticed that there are as many amusing results as there are accurate ones. And yet, although the technology does not work with 100 percent accuracy, it is good enough to be widely used. And even make you nervous.

NEC's Paul Howie says their face recognition system scans individuals for individual identifiers:

“For example, many consider the distance betweeneyes a unique characteristic. Or it may be the distance from the chin to the forehead and other components. We, in particular, take into account 15-20 factors that are considered important, as well as other factors that are no longer so significant. A three-dimensional image of the human head is created, so even if it is partially closed, we can still get an exact match. The system then takes the signature of the face and passes it through the database. ”

Is it worth worrying about face recognition software?

First of all, face recognition is data. Data can be collected and stored, often without permission. Once the information is collected and stored, it is open for hacking. Platforms with face recognition software have not yet been seriously hacked, but as technology spreads, your biometric data will be in the hands of more and more people.

There are also issues of ownership. Most people do not know that when they are registered in social media platforms like Facebook, their data from this moment belong to this very Facebook. Since the number of companies using facial recognition is constantly growing, very soon you don’t even have to upload your own photos to the Internet to be compromised. They are already stored there, and stored for a long time.

Speaking of software, they allwork differently, but basically they use similar methods and neural networks. Each face has many distinctive features (in the world it is impossible to find two identical faces, and yet how many were there in the whole history of mankind!). For example, the FaceIt software defines these features as nodal points. Each face contains about 80 nodal points, similar to those we mentioned before: the distance between the eyes, the width of the nose, the depth of the eye sockets, the shape of the chin, the length of the jaw. These points are measured and create a numeric code - a “fingerprint” - which then enters the database.

In the past, face recognition was based ontwo-dimensional images to compare or identify other two-dimensional images from the database. For greater efficiency and accuracy, the image had to be a face directly looking into the camera, with a slight dispersion of light and without much facial expression. Of course, it worked damn bad.

In most cases, images were not created in the right environment. Even a small play of light could reduce the efficiency of the system, leading to high failure rates.

3D has been replaced by 3D recognition. This newly emerging software trend uses a 3D model that provides highly accurate face recognition. Capturing a three-dimensional image of the surface of a person’s face in real time, the software identifies distinctive features — where rigid tissues and bone are most prominent, such as eye socket, nose, and chin curves — to identify a subject. These areas are unique and do not change over time.

Using depth and axis measurements that are notthe lighting affects, the 3D face recognition system can even be used in the dark and recognize objects from different angles (even in profile). Such software goes through several stages, identifying a person:

  • Detection: Take a picture using a digital scan of an existing photo (2D) or video to get a live picture of the subject (3D).
  • Centering: Having determined the face, the system marks the position of the head, size and posture.
  • Measurement: The system measures the curves on the face with millimeter accuracy and creates a pattern.
  • Representation: the system translates the pattern into a unique code. This code sets each pattern with a set of numbers representing features and facial features.
  • Matching: if the image is in 3D and the database containsthree-dimensional images, the comparison will take place without changing the image. But if the database consists of two-dimensional images, the three-dimensional image is decomposed into different components (as if two-dimensional images taken from different angles of the same facial features), and they are converted into 2D images. And then there is a match in the database.
  • Verification or identification: in the process of verification the picture is comparedwith only one snapshot in the database (1: 1). If the goal is identification, the snapshot is compared with all the snapshots in the database, which leads to a number of possible matches (1: N). This or that other method is applied as needed.

Where are face recognition systems used?

In the past, facial recognition systems foundthey are used mainly in law enforcement, because the authorities used them to search for random people in the crowd. Some government agencies also used similar systems for security and for eliminating electoral fraud.

However, there are many other situations in whichsuch software is becoming popular. Systems become cheaper, their distribution is growing. Now they are compatible with cameras and computers that are used by banks and airports. Travel agencies are working on the “seasoned traveler” program: with its help, they conduct a quick safety screening for passengers who voluntarily provide information. Airport queues will advance faster if people go through a facial recognition system that matches faces with an internal database.

Other potential uses includeATMs and cash dispensers. The software can quickly check the face of the client. After client's permission, the ATM or terminal takes a snapshot of the face. The software creates an imprint of the person who protects the client from identity theft and fraudulent transactions - an ATM simply will not give out money to a person with another person. Even the PIN code is not required.

Magic? Technology!

Development can be especially important and interesting.face recognition technology in the field of bank transfers. Recently, the Russian bank Otkrytie presented its own unique solution developed under the technology brand Open Garage: transferring money from a photo in the Otkrytie.Transvody mobile application. Instead of driving in the number of a card or phone, you just need to take a picture of a person who needs to make a transfer. The facial recognition system compares the photo with the reference (done when the bank issues a card) and prompts for the first and last name. You only need to select a card and enter the amount. Most importantly, customers of third-party banks can also use this function for transfers to customers of Otkritie - the sender of transfers can use a card of any Russian bank.

"Using customer photos instead of numbersa bank card is a fundamentally new approach to online transfers based on the use of a neural network facial recognition system that allows you to identify a client with a high degree of accuracy using his biometric data, says Alexey Matveyev, the head of the Department for Development of Partner Systems at Otkritie Bank - Service opens up completely new life scenarios for users to make money transfers. Currently, none of the participants in the financial market in the world offers this service to its customers. ”

Mobile application “Opening. Translations ”can be downloaded here.