Karl Martin expected to be long-finished his PhD by now. But what started almost seven years ago as a pure research project in the Department of Electrical and Computer Engineering at the University of Toronto has transformed into a commercial venture set to change how we interact with the emerging "surveillance society." The project promises something that most new technology in this everybody-knows-your-business age doesn't even bother trying: improving personal privacy. As surveillance cameras pop up in increasing numbers in private and public spaces, Martin's efforts show that, with the right design, we don't have to sacrifice privacy at the altar of security and convenience.

"It's been exciting to be involved in making something real out of the research I've been doing," says Martin. "The research timeline is usually so slow, it's been surprising how quickly this has moved along."

Martin started his PhD back in 2003, working under professor Kostas Plataniotis on image-processing encryption and compression techniques. As the project progressed, Martin wondered what kind of application the work might have and realized they might be able to create a privacy-protection system for video-surveillance cameras. Software would block faces and identifying features from live video feeds but that blocked information would still be available in the data stream in encrypted form. Identifying personal details could be retrieved if they were deemed important, say, if a crime was committed and the police needed to review the video. The rest of the time, privacy would be the rule. Why, if they don't have to, should security guards see transit passengers picking their noses or hotel guests bringing company back to their rooms?

"It's an excellent solution because there can be different levels of encryption depending on who is using the information and for what. The information is still there.  A TTC employee could see one thing. The police or a judge could see more," says Plataniotis. "If you look at what Google [which photographs city streets for its Street View project] is trying, with not much success I might add, they blur the information and destroy the original image. You can't recover it."

The work remained purely academic until the surveillance-camera issue exploded in 2007 when the Toronto Transit Commission (TTC) announced it would install 12,000 cameras across its network of subways, buses and streetcars. The U.K.-based group Privacy International filed a complaint to the office of Ontario Information and Privacy Commissioner Ann Cavoukian, calling the TTC plan "an inappropriate and poorly considered use of resources" that violates Ontario's Privacy Act (pdf). Cavoukian declared the TTC plan was within the law, but she didn't stop there. An advocate of "privacy by design"� a belief that technology should be able to provide convenience and security without sticking its nose into our private lives �Cavoukian prodded the TTC to do better. She had learned of Martin and Plataniotis's work and put them in contact with the transit authorities.

The result was a pilot project at the Bay subway station where the researchers have been collecting real-life video to experiment with. Although much of the work involves programming algorithms into computers and working out formulas with pencil and paper, the TTC partnership allowed Martin and Plataniotis to use themselves as guinea pigs (you can see a demo of the system here).

Meanwhile, the U of T's Innovations and Partnerships Office came on board. Working with the MaRS Centre, the office brings the ideas of students and staff to the marketplace. They help researchers patent their work, find public funding opportunities and seek private-sector partnerships. The office sees about 150 inventions each year and moves forward with between ten and 20 of them. So far, they've helped Martin and Plataniotis patent the idea, set up a company and partner with Burlington-based Xiris Automation. The project was also able to secure a $110,000 research grant from the Natural Sciences and Engineering Research Council of Canada that will help them build a prototype.

"It's critical to get that kind of funding at an early stage when it's too early to get angel or venture capital funding. That's why they call it the valley of death," says Lino DeFacendis, director of commercialization and business development at the innovation office.

The ultimate goal, foreseeable in the next year or two, is to bring to market surveillance cameras with the privacy-enhancing technology integrated into them. DeFacendis figures that public agencies in charge of transit systems and public spaces would be most interested, but purchases from banks, hotels and retailers is possible if the price point was right. There is no deal yet with the TTC.

Martin admits that people don't seem to care much about threats to their personal privacy. But he thinks that indifference will dry up as surveillance technology comes more and more intrusive.

"The public isn't yet aware how much they're captured on video and what can be done with that," says Martin. "It's not just who is monitoring it, but what happens to the information afterwards. Trends can be tracked over time, patterns detected. There will be a tipping point when having this technology on your surveillance system becomes a selling point."

Paul Gallant is a Toronto-based freelance writer who lives in the emerging Brockton Triangle neighbourhood.