« Smart Traffic Signals in Chicago | Main | Testing Facial Recognition in Japan »

April 23, 2006

NY Times Mischaracterizes Smart Cameras

From The New York Times:
A short article on the efficacy of surveillance cameras with a bit on smart cameras. First, it notes the debate over the efficacy of cameras. This is constant trope in stories with surveillance cameras with one side pointing to statistics in their city (such as Chicago) versus the studies in the UK that show cameras were generally not very effective. The end of the article is rather dismissive of smart camera technology:

Chicago is beginning a trial project using software that will sift through thousands of hours of video, trying to recognize unusual behavior, like leaving behind a suitcase. Such software is largely unproven, noted Elaine Newton, a fellow at the Stanford Law School Center for Internet and Society. "These things are going to have error rates," she said. Face recognition and other biometric applications are particularly difficult, and often the results depend on the quality of the image or the lighting. "Typically surveillance cameras are pretty low quality," she said. And they are often exposed to heat, which degrades image quality even more. As a result, Ms. Newton said, "real-time analysis of lots of cameras isn't something that's going to be invested in."

I don't with agree Elaine Newton as she is quoted in this article. First, Newton has done some work on smart camera topics with a paper on preserving privacy by de-identifying face images for the IEEE. However, she is totally wrong that real-time analysis is not going to be invested in. There are a whole slew of vendors selling these technologies, such as Vistascape, as well as government procurement policies that are seeking these technologies (Secure Border Initiative).

While there are significant problems with some smart camera technologies, such as facial recognition, there are other smart camera technologies that already are useful. If if real-time analysis is not yet useful, why wouldn't there be investment in the technology? Maybe I am too close to the issue, but I see enormous promise by adding "smarts" to the processing of camera footage.

I don't want to be too hard on Newton, because I am sure her words are somewhat twisted. Maybe she is saying that we won't add real time analysis to existing old analog cameras that don't have the resolution for facial recognition or analyzing crowd behavior. Nevertheless, this article doesn't accurately reflect what is happening with smart camera technology.

Posted by rshah at April 23, 2006 08:21 PM

Trackback Pings

TrackBack URL for this entry:
http://www.rajivshah.com/~rshah/directory/mt-tb.cgi/175

Comments

Post a comment