Berkeley professor looks to anonymize CCTV coverage

Those looking to prevent their children from looking at private images can use their television’s V-Chip. And if UC Berkeley professor Ken Goldberg has his way, those looking at public CCTV images will have to use his P-Chip, for privacy. Also known as respectful cameras, the video analytics method he is developing obscures the faces of people who appear on surveillance videos, although the white oval could be removed for law enforcement purposes. “Any technological measures that can be taken to mitigate the privacy invasion and avoid the chilling of legitimate conduct in public or private spaces that are being recorded is a good thing,” said Kevin Bankston of the Electronic Frontier Foundation.

We should not overstate Goldberg’s accomplishment, which relied on a camera system developed by the National Science Foundation-funded Team for Research in Ubiquitous Secure Technologies. “In its current state of development, the camera is only able to obscure the faces of people who are wearing a marker, in the form of a yellow hat or a green vest,” Technology Review reported. “The researchers use a statistical classification approach called adaptive boosting to train the system … but they also combined this classifier with a tracker, which takes into account the subject’s velocity, along with other interframe information.” Tests at a construction site resulted in correct identification of the colored marker 93 percent of the time. For now, Goldberg’s system relies on those entering the surveillance areas actively choosing a determining marker. The next step is making them smaller, perhaps to the size of a button.