NVIDIA is one of the leaders in the field of creating self-controlled cars and is also actively working on the development of artificial intelligence technologies. According to the company, video surveillance can be the next step in creating "smart" cities. On the eve of the annual conference of GPU Technology Conference, she unveiled a new platform for video analytics NVIDIA Metropolis. The idea is to use machine learning to analyze the huge amount of data received from CCTV cameras to ensure public safety, traffic management and resource optimization.

According to NVIDIA, hundreds of millions of CCTV cameras are used around the world, and by 2020 their number is expected to reach 1 billion mark. It is difficult for a person to cope with the analysis of such a volume of data. Therefore, most records are stored on hard disks for later study. NVIDIA believes that in-depth training in artificial intelligence can be used for video analytics with much higher accuracy than people can do. The computer is able to monitor the situation in real time. NVIDIA already cooperates with more than 50 manufacturers of CCTV cameras, including Hikvision.

"The advantage of deep GPU training is that the data can be analyzed quickly and accurately to get deeper into the bottom," said Shilang Pu, president of the Hikvision Research Institute in China.

A city with artificial intelligence systems connected to the cloud could find missing persons, notify residents about emergencies nearby, warn the police about the crimes committed or even send a signal about traffic accidents. Such a system is also capable of tracking people's behavior and collecting personal data for advertisers.