Researched and developed at the University of Toronto, secure visual object coding address concerns about privacy infringement by video surveillance. Nearly ready for commercialization, the technology has been tested with the assistance of the TTC and the backing of Ontario’s Privacy Commissioner.
Karl Martin isn’t abashed to show off surveillance camera footage of himself sauntering through an underground Toronto Transit Commission (TTC) corridor.
The president of newly-minted Toronto firm KMKP Engineering is twinned on side-by-side video screens playing back the same footage.
Martin and a colleague wander into the frame, stop and point back at the camera as they shuffle about. On the right side, the video looks normal and Martin’s face clearly identifiable. On the left side, nothing but pixilated garbage takes the place of his head.
Developed at the University of Toronto’s multimedia lab, this video encryption technology has exited the research phase and nears commercialization.
One of its first customers is likely to be the TTC, as the public organization rolls out video surveillance on its surface and subway vehicles and platforms. That pilot project comes thanks to the not-so-gentle prodding of Ontario Privacy Commissioner Ann Cavoukian.
“A camera doesn’t discriminate — it captures whatever is in front of it. It’s a dumb device,” Martin says at Cavoukian’s Privacy by Design conference Jan. 28.
Right now, he says, there’s no commercial tool for surveillance that offers privacy without impeding security.
Martin hopes to change that soon enough, and the TTC is likely to be the first place this technology is implemented.
When the transit commission for Canada’s most populous city announced it was going to expand video surveillance in 2007, it was investigated by the province’s privacy commissioner. Prompted by a complaint from U.K.-based Privacy International, Cavoukian delivered a report with several suggestions on how TTC could better ensure individual privacy rights were protected.
Cavoukian heralded Martin’s work in her report, suggesting it could ensure privacy, while still providing reliable evidence to law enforcement officials.
“It codes information in a way that it can be retrieved later on, if needed by the police,” Cavoukian says in an interview. “The beauty of it is that the footage most people would see for administrative purposes, wouldn’t reveal any personally identifiable features whatsoever.”
The 2007 report was recognized internationally as setting the standard for video surveillance privacy. Indiana University School of Law professor Fred Cate recommended the report to the U.S. Department of Homeland Security for use as a best-practices framework. Other industry analysts also recognized it as a step forward.
The TTC has adopted the recommendations, according to Michael Atlas, senior solicitor at TTC. That includes working with the University of Toronto researchers to develop their software.
The TTC has been installing more cameras over the past two years. The commission has followed other suggestions by Cavoukian, such as limiting the amount of time video footage is stored to a maximum of 72 hours for subways and 15 hours for surface vehicles.
When the police want footage from the TTC, two signatures must be present on the request form, including one from the chief of police.
“The TTC wants to use the least intrusive devices,” Atlas says. “The technology could also go one step further.”
It’s possible the system could also be programmed to identify a suspicious package left on a bus or subway, he says. “If it were left unattended for 10 minutes, it could alert you to that.”
Closed Circuit Television (CCTV) has long been a favoured method of police for nabbing bad guys, and deterrent to would-be vandals. But the concern over a trade-off of privacy rights for physical security has escalated in recent years. London, England has more than 10,000 publicly-installed cameras and 500,000 privately installed cameras that feed into the public system known as the “ring of steel.”
In 2007, New York announced plans to install 3,000 cameras in Lower Manhatten in an initiative known as the “veil”. More and more, Big Brother really is watching.
Martin’s video encryption technology negates the potential for any video voyeurism. Those monitoring the screens in real time would see a blurred-out or pixilated image in place of personally identifiable features. If a video needed to be reviewed by authorities for evidence purposes, they could then use an encryption key to view the footage as it was originally recorded.
The TTC has installed cameras on all of its surface vehicles. Subway “choke-points” are also covered off, Atlas says, and installations are still being done on subway platforms. New subway cars put on tracks this year will have cameras as well.
Martin plans to first release software that will process the image after it is recorded. Testing with the TTC video cameras will help determine where the software should reside in the system.
“We’ll allow them to do their testing,” Atlas says. “It would depend on whether they’d want to use it testing themselves, or using regular TTC patrons.”
Ideally, the encryption would be installed directly into the video camera so identifiable features were blurred by default.
The development team received a $110,000 research grant from the Natural Sciences and Engineering Research Council of Canada to help bring it from the drawing board to market. The TTC will be one partner in helping it get there, having already completed an offline pilot study of the technology (seen in video).
“We have ourselves as subjects in the subway to see how it would work in that type of environment,” Martin says. “We hope in the future to have a real time, live pilot study.”
When the time comes, you can bet that Martin won’t just be the man behind the camera. He’ll be putting himself in front of it, too.
Follow Brian Jackson on Twitter.