CMU develops surveillance system
Civil libertarians warn there is a darker side to emerging technologies, such as a new multiple-camera computer monitoring program Carnegie Mellon University researchers developed to track patients in a nursing home.
CMU researchers who developed the “Marauder's Map” program to predict the health of multiple nursing home patients based on around-the-clock observation of their movements are preparing to present their findings at an international conference on Computer Vision and Pattern Recognition in Portland, Ore.
While they concede their work may have implications for national security, and a CMU press release touted the use of such technologies “in airports, public facilities and other areas where security is a concern,” researchers say their focus was the health of the elderly.
“This had to do with health research for individual patients to alert their caregivers to subtle changes that could indicate health issues. It was not meant to be a nefarious Big Brother, big government-type project,” said research team leader Alexander Hauptmann, principal systems scientist in CMU's Computer Science Department.
Jennifer Lynch, a staff attorney with the Electronic Frontier Foundation said ultimately the system could pose serious implications for privacy and civil liberties.
“For example, if the government were to use this technology to track and identify individuals as they go about their daily lives, it would allow the government to create a digital dossier of the intimate details of a person's life. This creates a chilling effect on individuals' behavior,” Lynch said.
The development emerges in the heat of a national debate over government surveillance of people's phone records and Internet usage.
Hauptmann said the research is preliminary.
The team analyzed six-minute-long recordings from 15 cameras stationed around the home that ran continuously for 25 days.
“The next step is to run (the program) over 25 days of data,” said Shoou-I Yu. Yu, a doctoral student in CMU's Language Technologies Institute who deals in large data issues, came up with the idea to track multiple variables, including clothing color, movement trajectory, person detection and facial recognition to improve the accuracy of the tracking program.
Yi Yang, a post-doctoral student in CMU's Computer Science Department who rounded out the research team, said they developed formulas to allow them to combine the multiple variables to reach 90 percent accuracy.
Yu said he's never considered how others might use the new technology.
“I guess technology is always a double-edge sword,” he said.
Nita Farahany, a Duke University scholar who studies the legal and ethical implications of emerging technologies, said that's an important consideration.
“Every day, researchers reveal new technologies that enable more comprehensive surveillance of individuals. Surveillance can offer great benefits, such as improving health and safety for individual or enhancing productivity in the workplace. But comprehensive surveillance has a darker side.
“We should proceed carefully, establishing clear boundaries on the nature and extent of surveillance, and to encourage transparency of its use,” she said.
Debra Erdley is a staff writer for Trib Total Media. She can be reached at 412-320-7996 or firstname.lastname@example.org.