UA Team Wins Grant to Make Security Cameras Smarter

(PhysOrg.com) -- University of Arizona computer experts have received a federal grant to make security cameras smarter by programming software that helps the cameras detect suspicious behaviors, instead of just snapping images passively.

The UA team just received $2 million for two years, with the chance of getting another $3 million for an additional three years. The money comes from the Defense Advanced Research Projects Agency, known as DARPA. The team is led by Dr. Paul Cohen, Director of the UA School of Information Science, Technology and Arts. He also heads the UA Department of Computer Science.

“What we are trying to do is build a camera that can actually tell you that it sees particular activities,” said Cohen. “If somebody takes a package behind a car and emerges without a package, you’d like the camera to tell you that.”

The software would have to make expectations that something is not common behavior, said Cohen. At the same time, that software has to be careful not to make too many false positives.

“This thing has to be robust and it has to reason if it’s to avoid false positives,” said Cohen. “I think probably nobody would mind having security cameras, if the security cameras never made a mistake.” Cohen acknowledges there are privacy concerns and that the cameras could tag people or send alerts that aren’t correct.

Provided by University of Arizona