For example, when cash is being added to an automatic teller machine, two or more employees from a bank are required to do the job together, and if only one employee is doing it the camera detects the discrepancy and issues a warning to security staff.
Cameras can also recognize, for example, hand gestures that bank staff give signaling that a robbery is taking place.
Zhao says several state-owned banks have worked with the company for about six months, and in terms of criminal acts the slate remains clean.
After the camera system gathers information on human movement and behavior it analyzes the plethora of data and passes that information on to those who need it.
The system has commercial applications, too, being used, for example, in shopping malls to help analyze the movements of shoppers and what they are buying, and in museums to determine which exhibits attract the most attention.
Another system based on visual recognition that DeepGlint Technology has developed monitors motor vehicle traffic. It can sift through hours of video that shows thousands of vehicles to single out a motorist who has broken the law.
"If, for example, the police were hunting for a terrorist traveling in an identified car, our system can rapidly find it through a city's transport video," Zhao says.
Huang Yongzhen, an associate researcher of the Institute of Automation with the Chinese Academy of Sciences, says an intelligent camera system the institute devised has been used in Beijing's subway since 2007. It uses iris recognition technology and is aimed at preventing the theft of cables, particularly from places where it is highly dangerous to do so.
The technology is also used in coalmines. Because miners can be coated in coal dust, quickly determining whether all miners have surfaced can be problematic, and the iris recognition technology can do this quickly.
"In the practical application of image perception using artificial intelligence, China, with the United States, is at the forefront," Huang says. "In China, many digital technology experts in companies and academia are doing research in this field."
Within 10 years, he says, cameras will be able to gauge the behavior of a person at a distance of more than 10 meters in busy situations, such as in railway stations and airports.
|
|
|
|
|
|
|
|