As video surveillance increasingly moves to automated processes -- video analytics -- to detect suspicious activities, what becomes of the Surveillance Camera Players?
As video surveillance increasingly moves to automated processes -- video analytics -- to detect suspicious activities, here's a question that begs to be asked: What becomes of the Surveillance Camera Players?
Founded in 1996, this New York City-based political activist organization (Surveillance Camera Players) has inspired a worldwide movement of similar groups, all focused on doing the same thing: Performing in front of security cameras located in public places around the city. You can see an example of their work on YouTube (this is in a New York subway station and involves an interesting exchange with the police that probably wouldn't have occurred in London, where citizens have the right to get copies of video created by that city's ubiquitous surveillance camera system.) And here's a Spanish (and English) TV news report on the group: Dailymotion.com.
The Surveillance Camera Players perform to protest the loss of privacy from these cameras, but for all these years there has always been one underlying assumption: A human is watching. On the other side of the camera, in some distant (or nearby) security center, a security guard is watching a monitor on which the Surveillance Camera Players appear.
Increasingly, though, video analytics is replacing human monitoring. The assumption that there is someone on the other end to see the performance may no longer hold true. Of course, the whole purpose of analytics is to minimize the need for human monitoring -- not to eliminate it completely. After analytics has determined a suspicious event, the system -- whether built-in to the camera, or in a centralized processing system -- will send the brief video clip to a real live security person. This can save bandwidth, and enables a single security guard to effectively monitor dozens or even hundreds of cameras, as opposed to the handful that could be watched simultaneously at the old-fashioned security desk. Over the years, the field of video analytics has improved tremendously, learning to ignore common false alarms. For example, a crude "tripwire" drawn along a fence surrounding an airport -- designed to detect someone attempting to cut through or climb over -- might set off a false alarm every time birds fly across or sit on the fence, but a more sophisticated version can distinguish between birds and people.
Clearly the Surveillance Camera Players are harmless to democratic society, and it could be argued that their social commentary provides a vital voice in the ongoing debate concerning technology and privacy. Which begs another question: If they pose no risk, and distract security guards from their real work of spotting genuinely threatening events, will some clever analytics someday include an algorithm especially designed -- just as the birds get ignored -- to ignore the Surveillance Camera Players? And when that day comes, will they keep performing? Is "performing" for a camera that only a computer "sees" a bit like a tree falling in the forest with no one to hear the sound?
(Written by Cliff Roth and published on Videsign's Video/Imaging DesignWire, 21 October 2009.)
By e-mail SCP@notbored.org