Central London has at least 500,000 CCTV cameras and on a typical day an individual can expect to have their image captured 300 times. But what if those cameras remembered exactly where you'd been that day, or during the past weeks or months? And what if it weren't just the streets that were being watched, but your office too?
Questions of whether such surveillance is an infringement on privacy or simply benign protection aside, the reality is such systems are already emerging. Some vendors are pushing to make corporate physical security systems as pervasive and as easy to operate as Google. 3VR is one of the companies that believes in the benign power of surveillance and specialises in video analytic software that can, among other things, remember individual faces and pull up any video they've appeared in. "This will be of interest to anyone who wants to stop bad things from happening," says Tim Ross, the company's co-founder.
While critics say the spread of this kind of technology could be part of a Big Brother nightmare, Ross says it will actually be better for individual privacy in the long run. He argues that if the authorities can pinpoint exactly the video they need, they'll end up spending less time watching the rest of what goes on. "The trend towards more camera coverage is somewhat inevitable — in a post 9/11 world, there's too much need or demand for that," says Ross. "We think search is the key to privacy, not something that infringes."
Booming market
Corporate interest in physical security systems goes back, inevitably, to the terrorist attacks of 11 September, 2001. "9/11 increased the emphasis on all forms of risk management related to the potential for hostile human activity — both cyberdefence and physical defence," says Jay Heiser, research vice-president with analyst group Gartner.
As well as paying more attention to things like off-site backup systems, companies have started investing in ways of better locking down who can do what on their premises. The two main ways of doing this are video surveillance and access-control systems like key cards, which can be tied into IT authentication systems and feature biometric identification technology.
If 9/11made companies start to think they should be investing in systems like those seen on the TV series 24, five years on a host of start-ups have sprung up to take advantage of that demand. "The turning point was 9/11. It drove a lot of interest in homeland security, a lot of government spending and increased spending from large, Fortune 500 corporations," says 3VR's Ross.
On a basic level, companies have started digitising their video surveillance systems, which today are overwhelmingly analogue. That trend was in evidence when Cisco's acquired digital migration tools maker SyPixx. "Video is a good way to use up network bandwidth, and anything that does that fills Cisco's pockets, so why wouldn't they at least dabble in it?" says Gartner's Heiser.
Cameras that remember you
3VR has been drawing attention in recent months, partly because a successful funding round in January included investment from In-Q-Tel, a CIA-funded private venture firm. 3VR says it is also the first company of its kind to gain momentum in the mainstream business world, with customers including large companies like Morgan Stanley, GlaxoSmithKline, Lehman Bros and several large hotel chains.
3VR's system is designed to cut through the hours of undifferentiated video typically generated by surveillance systems and pick out the most relevant bits. At the heart of is the search engine, which Ross compares to Google, which indexes all the video as it is recorded, extracting all the data it can — including things like particular types of motion and face recognition — and putting it into a searchable database. The data is then used to trigger alerts around defined conditions, so that a feed will be brought to attention on the monitor when certain types of activity are detected. It also makes vast video archives easily searchable, Ross says.
One of the key data types is face recognition. Every time the system encounters what it thinks is a new face, it assigns an ID tag to the face, which it continues...
For more, click here...
