In Moscow, the face recognition system will work through CCTV cameras

Some of the 170,000 surveillance cameras in Moscow will be connected to the facial recognition system. The authorities recognized the successful two-month experiment with automatic recognition of faces, gender and age of people in real time (NTechLab Ltd. technology) - and now they are going to deploy the system to the whole city.

Face recognition system company NTechLab. Photo: NTechLab

Now the network of video surveillance unites access video cameras (95% of residential apartment buildings in the capital), cameras in the territory and in buildings of schools and kindergartens, at the MCC stations, stadiums, public transport stops and bus stations, in parks, underground passages, the official portal of the Mayor and Moscow government.

“The introduction of video analytics is a powerful driver for increasing the efficiency of both private and urban video surveillance systems. City residents have an additional level of protection, ”said Artyom Yermolaev, head of the Moscow IT Department. - Of course, all these opportunities should be implemented very responsibly. Our priority is a balance between confidentiality and security, and we follow a strict internal control policy that guarantees respect for the rights of citizens. ”

At present, over 3000 surveillance cameras installed at airports, railway stations, stadiums and other similar objects are connected to the facial recognition system. But in the future it will be expanded to other cameras , so that one of the world's largest video surveillance systems with face recognition will be created in Moscow.

About 16 thousand users have access to the city surveillance system - these are law enforcement officers, state and municipal organizations. Each of them has its own level of access, which "allows you to maintain the confidentiality of city residents"

Face recognition for people throughout the city provides unique opportunities for law enforcement officials. They can type in a photo of a person in the search - and find out where he is now, what route he takes, etc.

Identification example

NTechLab software uses neural networks and machine learning. It was this company that developed the famous FindFace application. At one time, the media covered several stories about this service, which quickly recognizes the identity of almost any passerby; it’s enough to take a photo of it and upload it to the app on a smartphone. The program was even used to de-anonymize Russian pornographic actresses — from the frames of the films, users found these girls on social networks, identifying themselves from school photos, etc.

By the way, law enforcement agencies can turn this focus now, but no longer on photos from social networks, but in the whole Moscow video surveillance system. The effectiveness of the NTechLab neural network is confirmed by independent comparative tests of facial recognition systems conducted by the US Department of Commerce and the University of Washington ( FaceScrub competition ). Accuracy of face recognition is about 70% . “Our technology is considered to be the best in the world in terms of the quality of work with the most complex of wild exploration images in the NIST metric. It includes photos of people taken in spontaneous conditions and with different resolutions, and the faces can be partially closed, ”explained the founder of NTechLab Artem Kukharenko.

But in video surveillance cameras, recognition accuracy is much lower: “Urban video surveillance cameras are dynamic: they move left and right, move in and out. In such conditions, even recognition of 60–70% of images is extremely difficult. The result of 30% is already cosmic, ” said Artyom Yermolaev in a comment for RBC.

Supervisory authorities should closely monitor that the system is not used for abuse, because there are certainly many who want to use this powerful tool for personal purposes: to track their wife, girlfriend or any other arbitrary person who has not committed any crime. Probably, access to face recognition should be strictly limited: only for specific single operations on a court order.

The Moscow Department of Information Technology noted that the introduction of the face recognition function has already increased the effectiveness of the investigation of offenses and the search for criminals. During the pilot tests with its help, they found and detained “more than 50% of lawbreakers who were searched using analytical algorithms” (6 people found, Bloomberg specifies ). Before that, some of them could not be found for many years.

The power of the system is not enough for total surveillance and recognition of all citizens in real time. That is, the system can not make surnames of all who enter each entrance of each house: this requires too many resources. So far, it can only find on all video cameras of individuals whose faces are loaded into the database. As soon as a person is detected, the system sends a notification to the law enforcement authorities.

Artyom Yermolaev said that the Moscow government now spends about 5 billion rubles ($ 86 million) per year to maintain the video surveillance system, and if all 170,000 cameras are connected to the facial recognition system, this amount will triple. Therefore, it is likely that only key cameras in certain areas will be selectively connected to the system.

Video from cameras is transmitted to a single data storage and processing center (ECDC) and stored for five days.

Moscow centralized video surveillance system - one of the largest in the world, and perhaps the largest. It can be compared with a centralized video tracking system for UK citizens. Although the exact figures are unknown, but according to the 2013 report, the British government had access to at least 70,000 video cameras. In Moscow, according to PwC, the density of video cameras is 54.6 units per square kilometer. And Moscow has become one of the first megacities that introduces face recognition in the urban video surveillance network.

“Video streaming analysis technologies are actively used within the framework of the“ smart city ”concept throughout the world, this market is constantly growing. Moscow was not just a trend, but proved to be a super-progressive capital - from the point of view of both scale and level of decisions, ”said General Director of NTechLab Mikhail Ivanov.


All Articles