New privacy-enhancing video surveillance technology developed by the University of Toronto may bring some peace in the tug-of-war between public safety and privacy concerns in mass transit systems - but also raises a fresh set of concerns.
The technology was endorsed by Ontario privacy commissioner Ann Cavoukian, who recently gave the Toronto Transit Commission (TTC) the green light to proceed with its plans to implement 11,000 video cameras by 2011 after investigating a privacy complaint lodged by UK-based NGO Privacy International. If implemented with strong protocols, this privacy-enhancing technology (PET) allows surveillance to be conducted without the usual concerns, says Cavoukian in her report.
Privacy by design
Developed by Karl Martin and Kostas Plataniotis, researchers at the faculty of engineering, their secure visual object coding application uses cryptography techniques to encrypt "objects of interest" within video frames -B faces or other features that may be used to identify a person - and store them separately. In order to view the original complete image, a decryption key is needed to restore the object of interest.
Activities could be monitored in real-time by security staff or retrieved from recordings but faces would be obscured.
"This is reversible, as someone with the proper authority who has the decryption key can retrieve the data if there's an incident. And it's exactly reversible once it's decrypted, so there's no issue in courts about image manipulation," says Martin.
The PET application is an add-on designed to work with a front-end video surveillance system that has object detection capabilities. "Facial detection has been around for a long time to varying degrees of performance. Many vendors already offer it in consumer applications like digital cameras so you can focus on faces when taking photos."
But incorporating facial detection in far more complex video surveillance systems is still in the early stages. There are still technology kinks to sort out in developing reliable facial detection. "Ultimately, whether our application will be useful depends on the accuracy of the front-end system's facial detector," warns Martin.
Martin and his team are partnering with the TTC on a research study to sort out kinks in their own application.
"There are practical issues in real-time operations, as the video comes in at certain frames per second, and we want to implement it so there are no lags. The TTC has provided an environment where we can test to see if it works."
The privacy-enhancing system requires significant processing horsepower, which will likely add costs to the $21 million price tag for the TTC's security system, he says. "Part of our work is to minimize that, but there's no mandate from the OPC that the TTC must implement our system."
But Cavoukian notes that surveillance systems should be designed upfront with privacy protections to avoid even costlier retrofits in the future.
While these technologies aren't ready for prime time yet, Martin believes they will evolve out of sheer necessity due to the huge numbers of cameras being implemented in public spaces everywhere. "The biggest problem now is there isn't enough manpower to monitor all these cameras being put up by the TTC and the police. All they can do is capture images so they can go back later if something happens."
One of two things will likely happen, he says. To reduce the operators needed to monitor video cameras live, these systems will be automated and designed with more intelligence so they can monitor faces, patterns and events, in addition to incorporating privacy protections - or a public backlash will occur. "It's possible the whole thing could collapse because there's a ridiculous number of cameras out there."
Quiet constituents
The TTC's plans for an expanded video surveillance system haven't roused much reaction from Torontonians, says Ed Drass, public transit advocate and columnist for commuter newspaper Metro News, which has a readership of about 400,000 TTC riders.
Although Drass often acts as a go-between, presenting responses by TTC officials to readers' questions and complaints, he says he's received only one e-mail about the proposed video surveillance system - and it was in favour. "Not much has been expressed to me either way," he says. "My conjecture is that riders are tolerant of it because they feel there's not enough security in the system."
Readers are less quiet on this score. For example, Drass has received many grumbles about the TTC's special constables, or lack thereof. "I've received several e-mails saying there aren't enough of them out there. In one e-mail, a reader asked if the special constables physically exist, or are they just in flyers and posters."
Public concern about security seems to come in waves, particularly after a crime occurs in the transit system, he says. "There was a spate of incidents involving bus drivers and passengers a few years ago, and I think this is where a lot of the energy for cameras got going. But the union was pushing against cameras because they might be used against bus drivers in disputes."
Drass says he has ambivalent views on the matter himself. "Cameras may act as a deterrent, but it's hard to tabulate that statistically. They may be useful for forensic purposes, but they don't stop anyone from doing anything - it's just a piece of technology sitting in the ceiling."
He says he's surprised there hasn't been more grass-roots opposition in Toronto. "This is as much an activist city as New York or London. But it's possible there's a tipping point in all this. If people suddenly start seeing cameras all over the streets and subways, that may be the point that opposition starts up - but I don't see reaching that tipping point soon."
Primordial fears lurking
A point that's often overlooked is that privacy legislation is ultimately about feelings, says David TS Fraser, a privacy lawyer at Halifax-based law firm McInnes Cooper. "Although the legislation is written in a way that talks about personally identifiable information and identity theft, it's ultimately designed to protect people's sensibilities about unwanted intrusions," he says.
PET technology may not be enough to address those sensibilities unless the rules governing the use of surveillance are stated. "While the technology may do a good job of limiting the actual intrusions, I'm not sure it does much to address people's feelings about being watched. Unless the policies and procedures around surveillance are clearly communicated, it won't diminish that visceral feeling of unease about being spied upon."
Fear of the unknown is at the core. "If you see a cop at a corner, you can tell from his uniform who he is, what he's looking at, and if you've aroused his suspicions," he says. "But a camera is completely faceless. You don't know who's watching and how the information captured is used - will it wind up on late-night television?"
He notes a significant number of videos in these shows displaying people caught in embarrassing situations come out of Britain, where an extensive network of cameras in public places is rousing a public backlash. Cavoukian noted in her report that U.K. camera operators have caught entertaining themselves by zooming in on attractive women. "If you're going to outsource surveillance to a bunch of badly-paid guys locked in dark rooms, they're going to see more bums than bombs," agrees Fraser.
He concedes that automating the enforcement of policies and procedures around surveillance with PET technology rather than relying on fallible human operators to refrain from misusing the information offers some comfort. But he warns this may have the unintended effect of increasing video surveillance. "Unfortunately, this stuff makes it more acceptable to put video cameras all over the place, and by making it better and safer with less intrusive technology, it may ironically lead to more surveillance."
Related content:
Privacy matters in the public eye
Five tips for wireless video: What you don't know could hurt you
Wi-Fi growth fuels video surveillance adoption
icon.
