Posted by Brian Brown on Thu, Dec 17, 2009
As a CCTV system operates, it collects loads of video data. But is the data being used to its fullest potential? Can we improve upon how this data is analyzed?
Currently in development, a next generation of smart CCTV system, known as Samurai, may be the answer. The article, "Smart CCTV learns to spot suspicious types", introduces the Samurai system as "capable of identifying and tracking individuals that act suspiciously in crowded public spaces." The article goes on to explain this suspicious behavior is determined by complex algorithms which profile individual's behavior in a given environment and compare it to the 'norm', such as mapping routine routes within an airport or busy train station. When suspect behavior is detected, an alert is issued to the human operator.
The human operator is key to the success of this system. Samurai relies on human feedback to adjust its algorithms. Take spotting a janitor for example, "an operator might reassure the system that the person with a mop appearing to loiter in a busy thoroughfare is no threat. When another person with a mop exhibits similar behavior, it will remember that this is not a situation that needs flagging up." The CCTV system learns from the feedback and recognizes it in the future...amazing.
According to members of the Samurai team, video analysis tools in use today follow rules which are 'too rigid' and they hope to improve upon this with Samurai. While initial testing has proved successful, the team will continue to hone their research into 2011. No expected release date was mentioned, but it's certainly interesting to see this development taking place.
Read Smart CCTV learns to spot suspicious types in its entirety here.
Join Accu-Tech's group on Linkedin.