MOSCOW IS SAFE
«I recommend that city governments and private organizations integrate facial recognition technology into their video surveillance systems to take advantage of the unprecedented opportunities it offers them. Thanks to our use of NtechLab’s facial recognition technology, it is now possible to create a cost-efficient, high-powered, centralized security management system for the city that analyzes large incoming data streams from city video monitoring cameras. This leaves us confident that the city’s residents have yet another layer of defense in their security. It goes without saying that these capabilities must be implemented very scrupulously. Our priority is to strike a balance between privacy and security. As such, we adhere to strict internal controls that ensure that citizens' rights are maintained».
THE PROJECT IN NUMBERS
The Department of Information Technologies (DIT) is the Moscow City Government agency that develops and implements state information technology and communications/telecommunications programs.
1. To create and implement a video analytics system to work with large sets of data to ensure public safety and to make life more comfortable for residents of Moscow.
2. To increase the storage time of data from city video monitoring cameras.
The approach to meeting these objectives utilized an algorithm developed by NtechLab to recognize faces in real time using neural networks and deep machine learning.
A one-of-a-kind video analytics solution was created in conjunction with the Department of Information Technologies (DIT) that integrated into the City of Moscow’s video surveillance system.
The solution’s features developed as part of the project completely automated operations to process video surveillance data for facial recognition in Moscow. Facial searches in video streams take less than 3 seconds. Working with video footage is now more accurate thanks to the system’s ability to select the best still frame from a group of images. The facial recognition system in Moscow finds and extracts important data from the entire video stream and saves that particular footage. This significantly decreases the amount of video data stored and considerably increases video storage time. The partnership between DIT and NtechLab resulted in speeding up police work with video footage to be 10 times faster.
No additional investments were needed for existing infrastructure (video cameras). The algorithm of the Moscow facial recognition system was customized to work with the video resolution typical of most of the city’s surveillance cameras (4CIF). The storage time for recorded video is now 6 times longer (from 5 days to 30). The number of criminals arrested has grown. Over 50% of those arrested had managed to evade detection for many years. Estimating the flow of people in public places makes it possible to tackle issues regarding urban planning, land development, and transport organization, along with large-scale events.
to a new level