With nearly 850 security cameras on campus, someone can always watch what students are doing. Soon, those cameras might be able to predict what they will do next.
The University installed new behavior-recognition software on 11 cameras on campus in August. The software is programmed to detect as many as 18 human behaviors that might indicate criminal behavior.
The program, Perceptrak, is being tested on six cameras lining the Washington Avenue Bridge and five cameras in the steam tunnels underneath campus, Wayne LaMusga, information technology specialist for the University’s Department of Central Security, said.
If the test goes well and the budget allows for it, Central Security officials will install the program on more cameras, LaMusga said. The next places that cameras are likely to be programmed are parking ramps and research buildings, he said.
However, some buildings, like Coffman Memorial Union, are too busy for the program to be effective, Steve Jorgenson, assistant director for the University’s Department of Central Security, said.
“The main test is the bridge,” LaMusga said. “The department wanted to test a highly trafficked area and a minimally trafficked area to determine the capabilities of the system.”
The steam tunnels will serve as a low-traffic test area, he said.
LaMusga said although the program is capable of monitoring 18 specific behaviors, not all of them, such as museum mode, are applicable to an area like the bridge.
The 11 cameras are set to detect three behaviors: converging people, lurking people and single person, Jorgenson said.
Converging people will alert a security monitor if a person’s personal bubble is invaded by others, Jorgenson said.
Lurking people is ideal for identifying potential robbers, especially in parking garages, because it can identify people who might be standing in wait.
Single person is set to notify a security monitor of a person walking alone late at night or early in the morning as a safety precaution.
However, the converging-people feature is not as practical during the day due to the large number of people who cross it, Jorgenson said.
The University decided to implement the new program after Central Security Director Bob Janoski visited Johns Hopkins University at Baltimore in March.
Installing and testing the University’s new software, as well as training staff to use it, will cost $24,000. The computers, which are designed to run the software on as many as 16 cameras, cost $17,000 to $18,000 apiece, Jorgenson said.
Despite the cost, the system has helped decrease reportable crimes at Johns Hopkins, Investigations Coordinator for Security Services at Johns Hopkins University Steve Ossmus said.
Johns Hopkins has 120 cameras, with 104 of them using the Perceptrak program, Ossmus said.
There were 23 reportable crimes that resulted in arrests on campus in 2005. There were 18 in 2006 and six in 2007 so far, he said.
“We attribute (the decline) to the cameras,” he said. “They take the place of 104 police officers around campus.”
Johns Hopkins also has added signs on campus that warn potential criminals that they are being watched, he said.
While campus crime has decreased, Ossmus said crime has increased outside the cameras’ reach.
“Crime is like a water balloon,” he said. “If you squeeze the balloon, the water is going to go somewhere else. Unfortunately, our concern is our community.”
Feeling safe at the University, however, is important to students, especially after the sun has set.
“(The cameras) make me feel safer because I have night classes that I cross the bridge for,” English sophomore Rachelle Cordova said.
How students will react to additional invasion of their privacy is yet to be seen, but biology junior Jon Malepsy said it’s not an issue for him.
“I don’t do anything too private on the bridge,” Malepsy said, “so the cameras are fine with me.”