How do neural networks track us?

A large-scale face recognition system has been operating in Moscow since 2019: 200 thousand cameras, and in 2020 it was deployed on a permanent basis

According to some data, ₽1,2 billion was allocated for this system, while according to other sources – “only” $3,2 million, which is six times less. A significant part of the amount will go to the monitoring system for participants in mass events, that is, rallies. What is it – concern for our safety or total control? We figure it out together with Nikolai Dubinin, the host of our YouTube channel Industry 4.0.

With the help of recognition systems, you can find and identify any person. Most often they are used on the streets, in public transport, public places – to monitor the situation and catch criminals.

Police also uses facial recognition cameras to search for missing people or identify people who need help.

On the roads recognition cameras help to identify violators and even send them a fine.

In medicine analysis of the patient’s face helps to more accurately determine the symptoms and select the optimal treatment.

In business centers, offices and sensitive facilities such systems are used to control and restrict access, instead of security and electronic passes.

In shops and cafes the system recognizes the client, helps to choose the right one and pay by person. Such technologies are already used by KFC and Alibaba. The system can also transfer information about all visitors to the phone to the seller or security guard. So the staff will know if a VIP client has come to them or, conversely, a potential thief and intruder.

In state institutions and banks recognition systems are used to register and issue documents. They also replace the passport and other documents in order to withdraw money from an ATM, get the necessary service or loan. In this case, the system itself will check you on all available databases and make a decision.

Phones and other gadgets face recognition systems replace the login and password for access. They can also be used in the smart home system: so that it opens the door for you or turns on the light upon arrival.

Facial recognition is a technology based on artificial intelligence. A special algorithm processes a photo or video, finds a person’s face on it and compares it with those that he has in the database. It then converts the face into a set of digital data called biometrics.

Biometrics is used for identification or authentication. Identification helps to determine who is in the photo or video. Authentication – to confirm that this person belongs to a specific person. Fingerprint recognition works the same way.

The largest face recognition system is in China. At the end of 2018, there were 200 million surveillance cameras, and in 2020 more than 600 million are predicted. Five Chinese cities at once are in the top 10 cities in the world in terms of the number of surveillance cameras: Chongqing, Shenzhen, Shanghai, Tianjin and Jinan.

With the help of cameras, the Chinese government monitors 2,5 million Uyghurs, a national minority that professes Islam. The country also has a social rating system. The higher it is, the more opportunities you have: flights abroad, a profitable mortgage, a prestigious job. This rating is calculated based on data received from surveillance cameras.

If in China face recognition is deployed at the state level, then in the US it is a global market for commercial projects. In a number of US states – for example, in California – facial recognition was banned, but only for the police and special services.

But the technology is actively used by social networks, online services and mobile applications. Facebook, Google, Apple, Microsoft, IBM, Amazon have their own facial recognition algorithms in the USA. Moreover, the first two are among the best in the world: with an accuracy of up to 99%.

Since 2018, the EU has been implementing the GDPR – General Data Protection Regulation. It does not allow monitoring the movements of people without their express consent. This is associated with the fact that in Europe face recognition technologies are developing much more slowly than in other regions.

In our country, in addition to the surveillance system on the streets of Moscow, face recognition is widely used by the Bank in our country. Since 2017, a nationwide program has been in place to collect faces, voices, irises and fingerprints.

Since 2006, we have had a law on the protection of personal data: you must ask for permission to collect and process data, and you can store them only within the country. But, apparently, this does not apply to government agencies.

The Russian NTechLab competes most successfully with the American giants – it is she who is behind the surveillance systems in the Moscow metro. She also owns the popular FindFace app.

With its help, earlier it was possible to search for people in social networks by photo. Among them were porn actors, and then criminals.

Other big players include VisionLabs, and Sensemaking Lab. They also make products for city and federal services, cooperate with Russian and foreign companies: retail, mobile operators, banks. VisionLabs, in particular, is the second major contractor in Moscow, a developer of solutions for public transport.

This article by Gemalto.com outlines the top seven facial recognition trends in 2020. For example, the race of leaders in the field of biometric technologies: Google, Apple, Facebook, Amazon and Microsoft (GAFAM). And also – Deep Learning – deep learning: when a neural network is trained on an increasing number of models, becoming better and more accurate. Another interesting trend is emotion recognition, a technology that is increasingly being used in marketing.

Subscribe and follow us on Yandex.Zen — technology, innovation, economics, education and sharing in one channel.