Engineers Test Highly Accurate Face Recognition

The work of postdoctoral researcher Allen Yang, of Professor Shankar Sastry's Heterogeneous Sensor Network (HSN) group at the University of California, Berkeley, is the subject of an article in Wired magazine where a new facial-recognition algorithm was created by Yang with the help of researchers at both UC Berkeley and the University of Illinois at Urbana-Champaign.

"Most algorithms use what's known as meaningful facial features to recognize people-things like the eyes, nose and mouth," says Dr. Yang. "But that's incredibly limiting because you're only looking at pixels from a designated portion of the face and those pixels end up being much smaller than the whole image. Our algorithm shows that you only need to randomly select pixels from anywhere on the face. If you select enough of them, you can produce extremely high accuracy."

Yang's new algorithm may signal a quantum leap in face-recognition technology. Professor Ssstry, dean of UC Berkeley's College of Engineering notes that Yang's new method obsolesces years of research in this field.

Nonetheless, the new technique could have profound impact in many areas, with new models for online advertising, new ways of annotating video and still images, and new techniques for identifying people in public places.

See the complete article in Wired.

Debugging Election Codes

An announcement on UC Berkeley's Electrical Engineering and Computer Sciences website tells of an article featuring David Wagner in the March issue of a Berkeley Engineering publication about his work reviewing voting machine systems code.

Professor Wagner, as the Principal Investigator of a joint UC Berkeley-UC Davis project commissioned by California Secretary of State Debra Bowen, led a team whose comprehensive examination found major vulnerabilities in voting machine systems.

While the machines were questioned immediately by grassroots activists, mainstream politics and media viewed their concerns about voting machine security as mere lunatic fringe behavior. However, according to Wagner, forward-thinking election officials changed this opinion. "Some elections officers took the activists' concerns seriously and forced these vendors to pry open the covers and hand over the source code," Wagner recalls. "That's what made it real; we could actually examine the code, so it wasn't just speculation anymore."

While Wagner's review prompted Bowen to limit the machines to one per polling place, a well-designed electronic voting machine could be a benefit to democracy.

See details in Innovations.

Ranking Corporate America on Identity Theft

The New York Times covered a report compiled by Chris Hoofnagle at the Berkeley Center for Law and Technology at the University of California at Berkeley on the institutions most frequently cited by consumers in fraud complaints.

The country's largest banks and phone companies showed up most frequently, of course. To account for size, Mr. Hoofnagle factored in the total amount of deposits per institution as of Dec. 31, 2006.

Mr. Hoofnagle said he believe the study was an important step in creating an "identity theft marketplace" for consumers.

"I've been working for years to try to spark a market, a true market, for competition on preventing fraud," he said. "Some of these institutions have attempted to compete based on advertisements, but I'm a real believer in the idea that if you give consumers information, they can make better decisions."

For the complete report, see Measuring Identity Theft at Top Banks.