Carnegie Mellon Researchers Release Details of Automatic Video Surveillance Tech

Declan McCullagh of CNET reports, "Computer software programmed to detect and report illicit behavior could eventually replace the fallible humans who monitor surveillance cameras. The U.S. government has funded the development of so-called automatic video surveillance technology by a pair of Carnegie Mellon University researchers who disclosed details about their work this week -- including that it has an ultimate goal of predicting what people will do in the future. 'The main applications are in video surveillance, both civil and military,' Alessandro Oltramari, a postdoctoral researcher at Carnegie Mellon who has a Ph.D. from Italy's University of Trento, told CNET yesterday."

He goes on, "Oltramari and fellow researcher Christian Lebiere say automatic video surveillance can monitor camera feeds for suspicious activities like someone at an airport or bus station abandoning a bag for more than a few minutes. 'In this specific case, the goal for our system would have been to detect the anomalous behavior,' Oltramari says. Think of it as a much, much smarter version of a red light camera: the unblinking eye of computer software that monitors dozens or even thousands of security camera feeds could catch illicit activities that human operators -- who are expensive and can be distracted or sleepy -- would miss. It could also, depending on how it's implemented, raise similar privacy and civil liberty concerns."

Read more here.

Image: Courtesy Flickr/ swanksalot