RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

The facial recognition system for 3.5 years helped identify about 1.5 thousand criminals in Moscow

Customers: Moscow Department of Information Technology (DIT)

Moscow; State and social structures

Product: FindFace

Project date: 2016/03  - 2018/09

Content

Moscow Smart City

Main article: Moscow Smart City

Video surveillance system in Moscow

Rostelecom provides the functioning of the video surveillance system of the city of Moscow PVN (49,797 cameras), DVN (10,111 cameras), MMS (736 cameras). The volume of work performed is about 50% of the total number of cameras installed.

The video surveillance system is built on the basis of the "Service Model"

  • The city orders video transmission services from telecom operators and pays for receiving video image from cameras to a single data storage center (ECCD)
  • ECCD is built by the city on a separate order and performs the functions of access to video cameras and data archives for users of the video surveillance system
  • The purchase of cameras, installation, configuration and subsequent operation of equipment is paid by the operator

Main article: How the city video surveillance system is arranged in Moscow

2024: The facial recognition system for 3.5 years helped identify about 1.5 thousand criminals in Moscow

The Sphere facial recognition system from the beginning of work (from September 1, 2020) to the beginning of February 2024 helped identify about 1.5 thousand dangerous criminals in Moscow, including murderers, rapists and robbers. About this in the metropolitan Department of Transport told on February 4, 2024.

As noted in the department, "Sphere" converts the face of a passenger who passed through the turnstile in the metro into a "unique biometric key" and checks with the databases of law enforcement agencies. If the system finds a match, it passes the information to the police. Deptrans emphasizes that the system "recognizes" only those people who are in the search databases, the rest of the passengers remain 100% anonymous - only law enforcement agencies have access to information.

Since the start of work, the Sphere facial recognition system has helped identify about 1.5 thousand dangerous criminals in Moscow

According to the deputy mayor of the capital for transport Maxim Liksutov, by the beginning of February 2024, thanks to the video system installed in the Moscow metro, a total of 1,324 people were found who were missing, including 280 children. In addition, with the help of the "Sphere" more than 9 thousand people were found on the federal wanted list.

Security expert Alexander Vlasov, commenting on the numbers of Deptrans Moscow in a conversation with BFM.ru, noted that the Sphere system is very correct, very good, and it only needs to be supported.

File:Aquote1.png
Every child found or every criminal caught is already good, otherwise you would have to spend a lot of money to either raise volunteers or raise police to look for these children. Therefore, we can say that the system works very well. It works simply: you upload photos of lost children or lost elderly relatives, wanted criminals, and video cameras compare those who pass. And if the data matches, they signal to the central point, "Vlasov said.[1]
File:Aquote2.png

2022

Facial recognition system in Moscow began to be used to search for draft dodgers

As it became known at the end of September 2022, the facial recognition system in Moscow began to be used to search for dodgers from partial mobilization. There are already the first detainees.

In particular, the Telegram channel Shot on September 28, 2022 reported that with the help of this system, the police detained four citizens, who were then sent to the military registration and enlistment offices. Channel "Caution, News" on September 29 reported another case in Moscow.

Facial recognition system in Moscow began to be used to search for draft dodgers

In theory, the facial recognition system can work to find mobilized, but not any. Most likely, we are talking about those who are hiding from the draft and against whom a criminal case has been initiated, "Kirill Stepanov, partner of the Kaminsky, Stepanov & Partners legal agency, told Kommersant.

In his opinion, in the event of a declaration of martial law in the Russian Federation, video surveillance and face recognition systems in cities can be used to ensure control over the population and establish the location of persons. The most likely development of events may be the lifting of demands for a court sanction to request information subject to the secrecy of correspondence, as well as the restriction of the right to inviolability of the home, believes Nikita Filippov, head of the De Jure bar bureau.

According to him, there is no fundamental ban on the use of video surveillance and face recognition systems for identifying and searching for people in the legislation. Considering that the search for mobilized citizens is carried out on the basis of the 144-FZ "On operational-search activities," the powers of the internal affairs bodies may cover such events, he added.

The source of the newspaper, familiar with the technical and legal nuances of the work of city video surveillance systems, claims that by the end of September 2022 there are no registers of deviators with a set of necessary data to search for them through the capital's facial recognition systems.

File:Aquote1.png
For police officers to gain authorized access to the metropolitan video surveillance system and be able to find a person with its help, it is necessary that this person be in the database of wanted persons, he says.[2]
File:Aquote2.png

Moscow Mayor's Office signed contracts for 800 million rubles to improve the facial recognition system with city cameras

As it became known at the end of August 2022, the mayor's office Moscow signed contracts totaling about 800 million rubles to improve the system face recognition with city cameras.

According to the BBC Russian Service, four technologies developed in Russia and Belarus are used in the capital's facial recognition system at once. One of the systems belongs to NtechLab - this became known at the beginning of the summer of 2022. The rest were developed by the Russian companies Tevian (FaceSDK system) and VisionLabs (Luna Platform), as well as the Russian-Belarusian company Synes (Kipod). Kipod was used to track protesters in Minsk in 2020, the newspaper reports.

Moscow City Hall signed contracts to improve the facial recognition system with city cameras

According to him, Luna Platform is used in Alfa Bank, Sberbank, Tinkoff Bank and in the Russian Post systems. In addition, Luna Platform supports Face Pay. The instructions for the platform say that it knows how to determine emotions, but there is no list.

Finally, the FaceSDK is also able to track emotions on people's faces. There is anger, fear, sadness, joy and surprise on the list. According to a special formula, the result of the work in the aggregate is calculated, explained in the Department of Information Technologies of Moscow.

The publication notes that by the end of August 2022, the Moscow mayor's office has almost 190 information systems that collect information about the city and its residents. For several years, work has been underway on a single machine learning system that will extract data from all resources and platforms of the Moscow government. So far, there is no single system at the federal level, but the Main Directorate of the Ministry of Internal Affairs has an "Accounting and Barrage System - M," which conducts an end-to-end search of arrays of other bases, including facial recognition system.[3]

2021

Moscow City Hall has allocated ₽4,2 billion for smart cameras. They recognize garbage, animals, road accidents and cars stopped at the Moscow Ring Road

On December 24, 2021, it became known about the creation of a new video surveillance system in Moscow, on which more than 4.2 billion rubles will be spent. The corresponding tender on the public procurement portal was posted by the Moscow Center for Traffic Management (DPC). Read more here.

Violators of the mask regime in the shopping center in Moscow will be tracked by cameras

The Moscow authorities ordered the city shopping center to integrate entrance video surveillance cameras with a single data storage center for the mayor's office. So the mayor's office wants to strengthen control over the observance of the mask regime. This became known on October 20, 2021. Read more here.

Moscow will spend 130 million rubles so that city video cameras can find accomplices of criminals

On September 2, 2021, it became known about the upcoming update of the automatic recording subsystem for video information indexing scenarios (PARSIV). The Moscow Department of Information Technology announced a tender for 130 million rubles, within the framework of which it is planned to create technology that allows city video cameras to calculate the routes of movement of criminals and their alleged accomplices. Read more here.

3 billion rubles allocated for video surveillance with face recognition in New Moscow

On January 25, 2021, it became known about the state procurement of the Department of Information Technologies (DIT) of Moscow, within the framework of which the department is ready to pay 2.917 billion rubles until 2025 for the development of a video surveillance system with the ability to recognize faces in the Novomoskovsk and Troitsky administrative districts. Read more here.

Moscow Mayor's Office obliged the shopping center to connect their cameras to the city video surveillance system

On October 20, 2021, it became known that the Moscow Mayor's Office obliged shopping centers (shopping centers) to connect their cameras to the city video surveillance system in order to monitor the observance of the mask regime of store visitors. Read more here.

In Moscow schools will be launched facial recognition system

On September 8, 2021, it became known about the plans face recognition of the Moscow authorities to introduce the system in the city's schools. It is assumed that schoolchildren and teachers will be allowed into educational institutions only after. biometric authentications

Writes about this Businessman"" with reference to the review of the head Department of Information Technologies (DIT) of Moscow Eduard Lysenko to the draft government decree, developed, Ministry of Digital Development which is proposed to expand the cases of collection and processing. biometrics The interlocutor of the publication, familiar with the situation, noted that the idea to introduce biometrics in schools was raised in the spring of 2021 at one of the of the Security Council of the Russian Federation meetings after the execution in. Kazan

Facial recognition system will be launched in Moscow schools

DIT said that by September 2021, the government is exploring the possibility of introducing a facial recognition system in schools, but "it is now premature to say that it will be introduced." DIT stressed that the system will be applied only by decision of the director, teachers and parents.

At the same time, there should be no legislative problems, experts interviewed by Kommersant believe. But technical difficulties may appear, because the collection of biometrics is regulated by clear regulations: data must be stored in certified processing centers, and face and voice casts must be collected in special crypto cabinets using secure software.

The DIT initiative is "an alarming trend in terms of personal data leakage," Ivan Begtin, chairman of the Association of Data Market Participants, told the newspaper. So far, fraudsters do not make masks with people's faces, but this is possible, he noted. In addition, it is not clear how face identification will help protect children, and school students may be responsible for the tragedies, Ivan Begtin said.[4]

Muscovites will be able to pay for travel on the bus through the face recognition system

Muscovites will be able to pay for travel in the city transport through the system. face recognition This became known on August 10, 2021 from the materials of the corresponding tender placed by the Center for Traffic Management (). DATA CENTER DATA CENTER LLC "" won the competition, Intelligent social systems which will receive about 329 million under the contract. rubles More. here

Maxima Telecom will install thousands of face recognition cameras in Moscow for 1 billion rubles

On July 23, 2021, it became known about the tender for the installation of 2 thousand video surveillance cameras on the roads of Moscow. The competition was won by "Maxima Telecom," which will implement the project for 1 billion rubles. The contract was concluded with the Center for Traffic Management (DPC) of Moscow. Read more here.

Screens with a facial recognition system for 1 billion rubles will be installed in the Moscow metro

On February 25, 2021, it became known about the purchase of digital displays equipped with video surveillance cameras for the Moscow Metro. Read more here.

The facial recognition system in the Moscow metro made it possible to triple the capture of people on the federal wanted list

The biometric identification system using video cameras in the Moscow metro made it possible to triple the number of caught persons on the federal wanted list, follows from the public report of the Deputy Head of the Department of Transport and Development of Road Transport Infrastructure of Moscow Alexander Garakoev.

The system was launched in September 2020, and by the beginning of February 2021, law enforcement officers, together with transport security officers on the territory of the Moscow metro, detained almost 800 people on the federal wanted list. On average, they identify and detain 5 people a day, Garakoev cited the data.

Slide from the presentation of Alexander Garakoev

The biometric identification system in the metro as of February uses more than 5 thousand video cameras installed on the entrance turnstiles. By applying a transport card to the turnstile, the passenger simultaneously "lights up" his face with a video camera. According to Garakoev, the method provides the maximum quality of facial recognition, taking into account the fact that someone puts on a hood or tries to turn away.

Police officers working in the metro are equipped with wrist tablets connected to a single dispatch center and providing immediate, within 3 seconds, transmission to the tablet of information about the recognition of a wanted citizen with an indication of the time and place of his identification. After that, response measures are immediately taken.

In addition, there is a web interface with a database containing information about the movement of the identified person: you can track his movement in 30 days, the Moscow Department of Transport official added.

Slide from the presentation of Alexander Garakoev

The Ministry of Internal Affairs has a database with images of persons with whom the system is checked. In 2020 , the DIT of Moscow cited data that the federal wanted list at that time contained more than 200 thousand photographs. And the entire metro video surveillance system includes a total of more than 56 thousand video cameras.

As for equipping the biometric system with other vehicles, GUP Mosgortrans is also introducing it at facilities, said Alexander Garakoev. In addition to bus stations, it is also equipped with rolling stock: 600 units of electric buses and about 470 trams. As of February 2021, this system is being tested at facilities.

Video cameras with a biometric identification system allow you to pay for travel on the subway by a person: if a passenger is connected to a bank card where there is his biometric data, he will pass with an open turnstile.

2020

Identifying about 200 quarantine violators

On March 19, 2020, it became known that about 200 quarantine violators were identified by the Safe City video surveillance system in the capital.

Such data are provided by the Main Directorate of the Ministry of Internal Affairs in Moscow. In addition, more than 4 thousand crimes were solved using the city video surveillance system. As of March 2020, more than 178 thousand cameras have been integrated into the Safe City agro-industrial complex.

The effectiveness of video surveillance systems in solving crimes gives an annual increase of 15-20%. It is planned that in 2020 more than 9 thousand CCTV cameras will be installed in the capital. Then the number of devices in the metropolis will reach 190 thousand[5].

Muscovites will be installed total surveillance during quarantine

The city authorities are going to launch in Moscow a system for tracking the movement of Muscovites during the quarantine world due to the coronavirus pandemic. The system will recognize people by their faces, which will help identify violators of the strict restrictive world established in the Russian capital.

According to Kommersant, the equipment for the system has been fully purchased, but its deployment will require additional time. The timing of the final launch of the surveillance system for Muscovites and guests of the capital has not been established - this will take several more weeks.

The authorities Moscow are acting on the example of a number of countries, Europe as well as Taiwan , China where algorithms for identifying persons are already used to search for quarantine violators. In this case, the system will be built on the basis of the company's technology. For Muscovites NtechLab[6] will be[7]

Main article: Coronavirus COVID-19 in Russia

Moscow cameras will recognize criminals by iris, voice and tattoos

Moscow cameras will recognize criminals by the iris of the eye, voice and tattoos. Danila Nikolaev, General Director of the Russian Biometric Society, told RBC about the development of the corresponding system. The Ministry of Internal Affairs confirmed this information to the publication. Read more here.

Sobyanin: the facial recognition system will fully work in the metro by September 1

Facial Recognition System will fully work in the Moscow metro by September 1, 2020. The capital's mayor Sergei Sobyanin said this during an expanded meeting of the GU board MINISTRY OF INTERNAL AFFAIRS Russia on. To Moscow 

According to the mayor, systems based on artificial intelligence were connected to HD-cameras of street surveillance. These measures will have to reduce crime in Moscow and recognize criminals "in a split second," Sobyanin said.

At an expanded meeting of the board of the Main Directorate Ministry of Internal Affairs Russia in Moscow Sobyanin noted that the system has already been massively introduced on the streets of the city from January 1, 2020
File:Aquote1.png
This is a different level of work, these are fundamentally different opportunities. We must work out the possibilities for video analytics, - said the mayor of Moscow.
File:Aquote2.png

According to him, the new system will also help identify the most criminogenic places in the city and collection points for illegal migrants. By the way, a quarter of crimes in the capital are committed by foreign citizens. In general, in recent years, there have been fewer murders, serious crimes, robberies, thefts and car thefts in Moscow.

Sergei Sobyanin noted that during 2019, the mayor's office tested the facial recognition system, which showed its effectiveness, and on January 1, 2020, the technology was introduced en masse and city video cameras were replaced with devices with support for HD-quality shooting.

In October 2019, Russian Interior Minister Vladimir Kolokoltsev said that the face recognition system has been tested since 2017 for 1.5 thousand cameras. The experiment was recognized as successful: the Ministry of Internal Affairs using the system managed to detain 200 criminals.

In January 2020, opposition politician Vladimir Milov and human rights activist Alain Popov filed with the capital's headquarters of the Ministry of Internal Affairs and the Department of Information Technologies (DIT). The complaint says that the facial recognition system in Moscow used by the authorities, including at rallies, is illegal, images of citizens from databases must be removed.[8]

Moscow buys 1.9 billion rubles of equipment for the face recognition system in transport

On January 17, 2020, it became known about the plans of the Moscow City Hall to purchase 1.9 billion rubles of equipment for the face recognition system in transport. The corresponding competition was announced by Electronic Moscow JSC.

According to Kommersant, citing public procurement data, it is planned to purchase 454 servers, each of which has two 18-core processors. Most of the servers must be in the "blade chassis" format - they take up half the space than ordinary ones, they are installed in special mounting cabinets.

In addition, the purchase indicates 50 computers with eight Nvidia Tesla T4 graphics accelerators, the power of which is equal to the performance of 30-40 ordinary computers, as well as a storage of 192 drives with a capacity of almost 1.5 petabytes (1.5 thousand terabytes).

The structures of the capital's mayor's office announced a tender for the supply of equipment for "video data processing" in transport for 1.9 billion rubles, ordering more than 450 servers and a data storage system with a capacity of almost 1.5 petabytes

The Moscow Department of Information Technology told the publication that the equipment is needed "to ensure the operation of the video surveillance transport segment," including in the metro and ground transport, and will help to even more effectively solve problems related to the operation of transport.

According to the newspaper, the Moscow government plans to create a "single video space" in which a powerful computer system will receive, store and analyze pictures from cameras installed in entrances, streets, train stations, transport, traffic lights and city highways.

According to Chief information officer Valentin Sokolov, given the use of an 8 TB SSD, it can be assumed that Moscow will use the system to analyze and predict car flows, track the situation on the streets of the city and notify "relevant services in automatic mode."

Roman Milyaev, senior manager of the IT consulting group at KPMG in Russia and the CIS, noted that the video surveillance system in public spaces in Moscow is the source of "a huge amount of video data - tens of gigabytes of traffic per second."[9]

2019

Maxima Telecom will supply servers for the facial recognition system in Moscow

On November 20, 2019, it became known that Maxima Telecom would provide the Moscow City Hall with servers for the facial recognition system. The company became the winner in the tender, offering a price of 1.15 billion rubles. Read more here.

All surveillance cameras in Moscow will recognize faces

On October 22, 2019, it became known that all CCTV cameras in Moscow will be equipped with a face recognition function. This was stated by the head of the Ministry of Internal Affairs of the Russian Federation Vladimir Kolokoltsev, speaking at the International Police Summit in Seoul.

File:Aquote1.png
The largest network, numbering about 3 thousand intelligent video cameras for face recognition, is deployed in Moscow. In the future, it is planned to equip all city cameras with this function, the total number of which exceeds 160 thousand, - he said (TASS cites a message from the press service of the Ministry of Internal Affairs).
File:Aquote2.png

All cameras in Moscow will be equipped with a video recognition system for faces

According to Kolokoltsev, the facial recognition system helps police officers investigate crimes more efficiently, and also simplifies "analysis and forecasting of the operational situation."

File:Aquote1.png
Modern police officers have a wide range of technologies and tools that our predecessors could not even dream of. It is possible to learn how to use these resources, develop new approaches and effectively fight crime today only on the basis of cooperation of law enforcement agencies with the use of achievements in the field of high technologies, - said the head of the Ministry of Internal Affairs.
File:Aquote2.png

He also recalled that Russian police successfully used technical capabilities to prevent crimes and acts of terrorism as part of security measures during the 2014 Winter Olympics and the 2018 World Cup.

CCTV cameras in Moscow are located metro, entrances, schools, kindergartens, stadiums and other crowded places. Earlier in 2019, it became known that with the help of facial recognition systems in the capital, about 300 people were detained in two years. So, since the beginning of the experiment in the metro (it began in March 2018), the technology has contributed to the detention of 62 people on the federal wanted list.

The court allowed the authorities to monitor people through facial recognition systems

In November 2019, the Savelovsky District Court of Moscow dismissed the claim of a resident of the capital of Alyona Popova, who demanded that the use of facial recognition system in the work of surveillance cameras be declared illegal. The court concluded that surveillance is not an interference in the privacy of citizens.

As Vedomosti writes with reference to the motivation part of the court's decision, the video surveillance system, although it uses recognition technologies, does not allow establishing a person's identity - for this, the Unified Data Storage Center (ECCD) lacks the necessary personal data ( iris, height, weight and other biometrics). From the images received, the identity is established when they are compared with the photographs provided by the police. Also, the identified matches do not exactly guarantee the identity (probability - 65%).

Illegal use of the facial recognition system in the operation of surveillance cameras

At the same time, the court's decision states that Interior Ministry officers use access to the ECCD "in order to fulfill their official duties," for example, to detect people who are wanted or under administrative supervision.

Alyona Popova argued that the processing of biometrics obtained using video surveillance without the written consent of a person violates the law on personal data, as well as the right to privacy. In April 2018, she held a single picket near the State Duma building, for which the Tverskoy District Court fined her 20 thousand rubles. He studied the recordings from video cameras that recorded Popova's face, her image on them was increased 32 times.

According to the law on personal data, biometric information can be used only if there is consent in writing (except for the cases described in the laws on defense, security and countering terrorism). From this, Alyona Popova concluded that the facial recognition system operating in the capital is illegal.[10]

The start of a tender worth 1.2 billion rubles. for the purchase of equipment for the facial recognition system

On November 8, 2019, it became known that the Moscow Government announced a tender for the supply of equipment for providing computing resources of the city video analytics system, information about which was posted on the public procurement website. The customer is the government-owned JSC Electronic Moscow. The initial maximum contract price is 1.2 billion rubles. Applications are accepted until November 11, 2019, the winner of the competition will be determined on November 18. Read more here.

All surveillance cameras in Moscow will recognize faces

On October 22, 2019, it became known that all CCTV cameras in Moscow will be equipped with a face recognition function. This was stated by the head of the Ministry of Internal Affairs of the Russian Federation Vladimir Kolokoltsev, speaking at the International Police Summit in Seoul. Read more here.

How many suspects helped to detain the facial recognition system in Moscow

On June 27, 2019, the results of the test operation of the facial recognition system became known. With the help of 1000 "smart" cameras installed outside the entrances of residential buildings, 90 people were detained.

In addition, as the representative of the Main Directorate of the Ministry of Internal Affairs for Moscow told Vedomosti, video analytics systems at several metro stations allow you to detain from 5 to 10 criminals every month, without increasing the police personnel.

The system compares persons from the video stream with the database of wanted persons, and in case of coincidence notifies nearby police officers. The experiment of recognition systems on the urban infrastructure started in 2017, and in the metro - in March 2018.

Moscow reported on the results of the introduction of a facial recognition system

The facial recognition system was also checked at 17 mass events: in two years, the police detained more than 150 people in this way, said Dmitry Golovin, head of the city video surveillance department of the capital's information technology department.

The Speech Technology Center (CST) and VisionLabs confirmed that they have provided their developments for a pilot launch in Moscow. In particular, the VisionLabs system helped detain about a hundred people in the subway. Ntechlab said they were participating in tests in the metro, but declined to comment on cooperation with DIT.

The tests continue, following the results, from three to five companies will work with the transport complex at the same time, the DIT added.

Mayor Sergei Sobyanin at the end of May 2019 announced that 200 thousand cameras would be involved in the city's facial recognition system. It will become one of the largest in the world.

By the end of June 2019, the Moscow video surveillance network consists of 160 thousand cameras located in the metro, entrances, schools, kindergartens, stadiums   and other crowded places. Also, the Ministry of Internal Affairs  began testing DVR cameras with facial recognition technology.[11]

Competition for a large-scale facial recognition system

In the near future, a large-scale face recognition system will be created in Moscow. As the mayor of the capital Sergei Sobyanin said in May 2019 at a meeting with Russian President Vladimir Putin on the development of technologies in the field of artificial intelligence, the facial recognition system will consist of more than 200 thousand video surveillance cameras.

The competition for the development of the project will be announced by the capital's authorities together with the Ministry of Internal Affairs in 2019, writes TASS. According to the mayor, the new system, covering more than two hundred thousand cameras, will become one of the largest in the world, and only Chinese developments can compete with it.

Sobyanin noted that last year the facial recognition system has already passed the test during the World Cup. In addition, experiments using CCTV cameras installed in the metro and entrances revealed dozens of wanted criminals.

Currently, there are about 167 thousand video surveillance cameras in the capital. Law enforcement officers, as well as state and municipal organizations (a total of 16 thousand people) have one or another level of access to them.

Video from city cameras is analyzed in real time. The neural network-based system tracks people's movements, identifies faces in the video stream, and compares search results with databases. Using the system, you can determine the sex and age of a person.

Facial recognition system in Moscow. How it is used and how it will develop

In 2019, Moscow plans to move from experimental operation of video surveillance with face recognition to its large-scale use, deploying a system module with this function on 105 thousand video streams, Dmitry Golovin, head of the city video surveillance department of DIT Moscow, told TAdviser in February. He also told what technologies will be used in it, and how the system is already used.

Whose faces are hunted by the system

Face recognition is used in mass and sporting events. These are international football matches and major city events. For example, the technology was used in Luzhniki at the World Cup, at the fan festival. As a result, it was possible to identify 98 people in the search bases, they say in the DIT of Moscow.

File:Aquote1.png
And at the CSKA Champions League match - Victoria Pilsen in 2018, for example, there were 4 detainees from the category of persons who, by a court decision, were prohibited from attending sports events. It is impossible not to miss such people for significant competitions without the use of such technologies, - notes Dmitry Golovin.
File:Aquote2.png

How Facial Recognition Technology Works

The video surveillance system works on a centralized basis. There is a single data storage and processing center (ECCD), where video streams flock from all 167 thousand city video surveillance cameras.

The face recognition function is implemented in the form of a separate video surveillance system module in the ECCD. If law enforcement agencies are faced with the task of finding criminals, they use their database of photographs, with which, if necessary, they compare faces that fall into camera lenses. If the images match, the police receive a signal about it. Similarly, the system works to detain those denied access to sporting events.

File:Aquote1.png
We can connect any city cameras to face recognition at the right time, depending on the tasks. However, there is a limit on the number of simultaneous flows - now it is about 1.5 thousand. We are preparing to create a full-scale facial recognition system for 105 thousand video streams in 2019, - says the head of the city video surveillance department of the DIT of Moscow.
File:Aquote2.png

How often the system is wrong

According to Golovin, the accuracy of recognition depends on many factors: weather conditions, camera location, etc. If it is necessary to identify a person from a photo from the search database and these photos of good quality, then the system correctly recognizes a person with a probability of 95-97%. Such results, for example, were obtained during the World Cup, according to DIT Moscow.

Which software is used

At the beginning of 2019, the recognition module runs on NtechLab software. According to Dmitry Golovin, DIT tested other solutions, they also showed good results. But with a full-scale deployment of facial recognition, the software vendor may be different.

File:Aquote1.png
In 2019, we plan to announce a competition and choose the best offer for the ratio of quality and price characteristics that meet our requirements, - says the representative of DIT.
File:Aquote2.png

He added that the city is receiving video analytics services from the service model. One of the reasons is to have the most efficient algorithms at the moment, and not to deploy the only technology at home for years.

Also, according to Dmitry Golovin, DIT has a permanent laboratory that tests new algorithms that appear on the market.

Regulatory regulation

Comparison of two photographs - taken by camera and photos from the database - is not regulated by regulatory acts, Dmitry Golovin notes, adding that legislation in the field of video surveillance, especially video analytics, is only being formed.

File:Aquote1.png
For our part, we are actively participating in its development and have already sent a number of proposals in the field of regulatory regulation to the responsible authorities, "he said.
File:Aquote2.png

At the same time, at the beginning of 2019, there are already laws and acts indirectly related to video surveillance. For example, the law on personal data, the Administrative Code, which affects the issues of photo and video recording of violations, etc.

Functional Development Plans

It is planned to develop the functionality of the system. The most important thing for DIT, says Golovin, is not just the introduction of new algorithms, but the automation of user activity, the integration of technologies into their business processes, so that video analytics free users from routine.

For example, earlier, in order to check the quality of cleaning the yard, specialists had to go with an inspection to the places. Thanks to the city's video surveillance system, they can already check the courtyard from their workplace, an example is given by a DIT representative. In the future, it is planned to connect video analytics so that the system itself recognizes, for example, overcrowded garbage cans, and gives the responsible signal.

File:Aquote1.png
We want to attract the maximum number of users to work with the system, make it simple and convenient, - explains Dmitry Golovin.
File:Aquote2.png

In addition, DIT plans to order the development of augmented reality glasses with facial recognition for the police. As of February, it was known that DIT was testing facial recognition technology with the transfer of information to augmented reality glasses[12].

Speaking about who determines plans for the development of these systems in Moscow, Dmitry Golovin says that these users who form development tasks are executive authorities and law enforcement agencies. As experts in this field, DITs can demonstrate the capabilities of technology to users, allow them to project video analytics solutions onto their processes, prompting them to various projects.

Ethical question

In Europe and the United States, the question of the ethics of using face recognition in video surveillance periodically arises - does it violate the right to privacy. In 2018, in Britain, for example, the first lawsuit was filed against the authorities because of the use of such a system on this very basis.

Dmitry Golovin says that this has not happened in Moscow yet. The system is used in public places and solely for security purposes, he notes.

File:Aquote1.png
People trust us. In 2018 alone, more than 27 thousand citizens applied for archives of records of the city video surveillance system. The townspeople realize its advantages, and use it to find the culprit in an accident or theft of a stroller, for example, - the representative of DIT cited examples.
File:Aquote2.png

2018: Sobyanin - on the deployment of video surveillance with face recognition

In early December 2018, it became known about the plans of the Moscow authorities to deploy a citywide facial recognition system that can calculate wanted criminals.

As Moscow Mayor Sergei Sobyanin said at a meeting on the prevention of offenses, a new IT system operating through city video cameras (by the beginning of December 2018, more than 162 thousand devices were connected in entrances, courtyards and streets) will be launched in 2019. 40% of devices have already been replaced with more advanced CCTV cameras in 2018.

In 2019, a citywide facial recognition system will be deployed in the capital to search for criminals

According to Kommersant, citing the press service of the Moscow Department of Information Technology, old cameras that support resolutions up to 704 x 576 pixels are replaced with HD equipment (1280 x 720 pixels). It is necessary to update about 100 thousand units of equipment.

New devices can be used to monitor citizens included in the entrances of residential buildings. For example, by analysis, you can understand that there are too many residents at the entrance and calculate "rubber" apartments in this way, the information technology department said.

Speaking about the launch of a citywide security system, Sergei Sobyanin noted that "this is a new level of security, this is a new level of technological equipment and law enforcement agencies, and in general city services."

In the budget of Moscow for 2019, 7.5 billion rubles were allocated for the city video surveillance system. Another 6 billion rubles will be provided to citywide data centers.

The head of the Main Directorate of the Ministry of Internal Affairs of the Russian Federation for Moscow, Oleg Baranov, announced the important role of the Safe City video recording system, the elements of which are installed in various parts of Moscow. The use of the system made it possible to solve 27 murders, 77 crimes related to causing grievous bodily harm, 165 robberies, more than 300 robberies, he said.[13]

2017: Start of pilot project

In September 2017, the Moscow mayor's office reported that more than 3 thousand video cameras of the city's video surveillance network were connected to a facial recognition system. At the time of launch, this is one of the world's largest security systems, where identity identification is used on such a scale, according to the DIT of Moscow.

The implemented recognition system can establish the identity of a person in the video, his gender and age. The mayor's office and DIT of Moscow do not specify which product is used for recognition and how much the project cost. According to Bloomberg, face recognition technology in Moscow was developed by the Russian startup N-Tech.Lab[14]. In 2016, he released the FindFace mobile application, which allows you to find a person on social networks by his photo.

Experiments with facial recognition technologies in city video surveillance Moscow conducted in 2016

The algorithm introduced into the city video surveillance system is based on the use of neural networks and analyzes video recordings from city cameras in real time. Individuals on records are scanned so that they can be compared to information in various databases if necessary - for example, in law enforcement photo databases when it comes to finding an offender.

The process of identifying the identity, according to the authorities, takes several seconds. If the algorithm finds a person whose face is loaded into the database, it will send an alert to law enforcement.

The analytical system can also help law enforcement agencies, when catching a criminal, build a route for his movement around the city. The system itself will select the necessary videos from different surveillance cameras, identifying the suspect in the video.

DIT says that during the pilot tests, more than 50% of percent of law violators who were searched for using analytical algorithms were found and detained. Before that, some of them could not be found for many years.

The Moscow authorities say that law enforcement officers can obtain the necessary data upon request within the framework of the current legislation, and employees of state institutions gain access to video cameras only from the territories and routes for which they are responsible. Each call to the tracking system is recorded.

In total, about 16 thousand users are connected to the city surveillance system as of September 2017 - law enforcement officers, state and municipal organizations. Each has its own level of access to maintain the confidentiality of information.

File:Aquote1.png
Residents of the city have an additional level of protection, - the words of the head of the Moscow DIT Artem Ermolaev are quoted in the message of the mayor's office. - Of course, all these opportunities should be implemented very responsibly. Our priority is a balance between privacy and security, and we have a strict internal control policy to ensure that citizens' rights are respected.
File:Aquote2.png

DIT told TAdviser that the project is not commented on, since this is only a technological experiment so far, and it is too early to talk about any conclusions.

According to the Moscow government, as of September, the capital's network consists of 160 thousand video cameras and covers 95 percent of the entrances of residential buildings. Access video cameras, cameras installed on the territory and in the buildings of schools and kindergartens, at MCC stations, stadiums, public transport stops and bus stations, as well as in parks are connected to a single system.

In addition, by June 2018, CCTV cameras were to appear in 25 underground pedestrian crossings of the capital: in particular, in underground passages not related to metro stations and under the jurisdiction of the GBU. "Gormost"

File:Aquote1.png
Since the beginning of October 2017, a video surveillance system capable of recognizing faces has been testing in St. Petersburg. The FindFace program, to which all outdoor CCTV cameras will be connected, will be able to analyze the image received from them in real time and identify offenders.
File:Aquote2.png

2016

First stage of testing

In December 2016, DIT Moscow talked about plans to launch the second stage of testing the system face recognition for city cameras in 2017 video surveillances[15] of the first stage of the experiment showed that so far it is too expensive to use such technology.

File:Aquote1.png
If you implement existing solutions across the city, it will require a tenfold increase in video surveillance costs. The technology will cost several billion rubles, which is absolutely unacceptable for us, - said Ermolaev.
File:Aquote2.png

The department then expected to reduce the cost of face recognition several times.

In addition, at the first stage of testing, questions arose about the accuracy of recognition: There are questions not only about the price, but also about the quality of recognition: it turned out to be extremely difficult to even achieve recognition of 60-70% of images. After the first stage of testing, Moscow was looking for technologies that would achieve more results on moving cameras.

N-Tech Lab talks with Moscow Government

In March 2016, it became known that the company N-Tech Lab, which developed FindFace technology for facial recognition, is close to concluding an agreement with the Moscow government on the use of its development for processing images from the capital's CCTV cameras. This was reported by the British newspaper The Guardian, who talked with the creators of the company - Artem Kukhnarenko and Alexander Kabakov[16].

150 thousand CCTV cameras have been installed in Moscow. The idea is that in the event of any incident, people in the appropriate place could be identified by the databases of wanted criminals, court files and profiles on social networks.

Notes