Facial Recognition System Begins Operation in Moscow Metro
The press service of the Moscow Metro has announced that, since March, cameras equipped with facial recognition technology have been tested at several metro stations. This was reported by Kommersant.
Equipment costing around 30 million rubles has been installed at several high-traffic stations. The press service declined to name the specific stations but noted that special attention is being given to stations located near stadiums that will host this year’s FIFA World Cup. “A different number of cameras have been installed at each location to ensure all entering and exiting passengers are captured, taking into account the design and size of each station,” the press service stated.
The purpose of the facial recognition system is to identify individuals who are on the federal wanted list. The software can recognize up to 20 faces per second. When it detects a match from the database, it sends the information to an on-duty police officer for verification. According to the press service, this system has already helped locate a “significant” number of people.
It is not yet known how long the testing phase will last. However, after testing is complete, the metro plans to deploy the facial recognition system on a permanent basis. “We want to offer all Russian IT companies with high-quality facial recognition systems the opportunity to collaborate with the metro,” the press service said.
The pilot project uses software from several Russian developers, including Vocord, whose systems are already installed at train stations in Adler and Ryazan, as well as at a stadium in Omsk. According to Vocord’s technical director, Alexey Kadeyshvili, similar experiments have been conducted in the Moscow Metro before, but a full-scale rollout has not yet happened. To implement the project, at a minimum, new cameras must be purchased and installed facing the flow of people at an angle of no more than 15%. The general surveillance cameras currently in the metro are not suitable for facial recognition systems. “If you try to use them for this purpose, the error rate can reach 30%. In other words, the system would miss every third person on the wanted list, and there would be many false detentions. These problems will only get worse as the wanted persons database grows,” Kadeyshvili explained. He added that standard overhead cameras installed on station ceilings are also ineffective, as they capture the tops of people’s heads rather than their faces.
It was also reported today that Moscow’s Domodedovo Airport plans to implement a facial recognition system. The airport has already successfully tested the “Videotech” program developed by the Russian company group “HighTech.”