Good CCTV cameras are extra eyes and ears that do not know fatigue. And modern machine learning technologies can make them even more effective. For example, the Nest company, dedicated to developments in the field of smart home, taught its camera to determine who is exactly at home.
According to Mehul Nariyaval, Nest's product manager, we are surrounded by many movements that do not matter. Take the same fan: the algorithm recognizes it as a movement, but it is not so important. Therefore, in the updated software of the camera, Nest engineers tried to teach him how to react, not to traffic, but to signals.
Working together with their colleagues at Google, the company's specialists developed an algorithm that can analyze up to 2.6 million frames per day and send notifications to detect anything suspicious. Like the human eye, the camera may not be 100% sure of what it sees. But if she sees him as a suspect, she will tell you.
The machine learning process is still far from perfect, but, nevertheless, computers become smarter when doing their job. As Nariyvala said, if the computers of the past could only say "yes" or "no", now machines can determine a lot of "shades of gray."
In addition, the updated Nest Cam software stores information from the last 10 days (26 million frames) and gives you the opportunity to "scroll" in a matter of seconds. You can see the most relevant moments during the day. The program "takes" them out of a large amount of video information so that the user can, for example, see the details of the hijacking of their bike a few days ago.
The publication Nest video camera is getting smarter appeared first in Indocent.