Research aims to improve airport security

Published 25 January 2010

From body-part censors to cameras that recognize faces, Carnegie Mellon’s CyLab is working with security technology that assuages privacy concerns; CMU’s Instinctive Computing Lab, eventually envisions a system that can wipe out the body image entirely, picking up only weapons, which will appear to be floating in space

The underwear bomber’s thwarted attempt to blow up a Christmas Day flight jeopardized President Barack Obama’s plans to close Guantanamo Bay prison, refueled advocacy concerns about passenger privacy, and forced Yang Cai to revisit an algorithm that identifies the curvature of a female breast.

Sitting in a Carnegie Mellon University office surrounded by plaster of Paris molds of the human figure, Dr. Cai explained how his algorithm could put to rest qualms about overexposure from 3-D body scanners at security checkpoints. His algorithms measure the human frame and identify the breast and genitalia areas. The system then automatically blurs those regions, blacks them out with a bar or replaces the features with dummy human parts. “Every man could be Arnold Schwarzenegger, and every woman could be Marilyn Monroe,” he said.

Airport security, however, has never been a place for fantasy. Pittsburgh Post-Gazette’s Erich Schwartzel writes> that the holiday scare and recent news that the Transportation Security Administration (TSA) plans to deploy 450 full-body scanners nationwide have government officials planning for legislation and passengers preparing for lines.

 

Privacy questions

Everyone is worried about who’s looking at what. Two researchers at Carnegie Mellon’s CyLab are working with security technology that assuages privacy concerns. Cai’s research offers an automatic censor for certain body parts and the CyLab Biometrics Center’s programs recognize faces and irises.

 

Despite the fact that Cai’s system blocks out private parts, it will identify an object hidden around those regions.

Cai, who founded CMU’s Instinctive Computing Lab, eventually envisions a system that can wipe out the body image entirely, picking up only weapons, which will appear to be floating in space.

His research, started in 2000 and published six years later, uses “intrinsic landmarks” to identify the body’s regions of interest. The unit of measure is the human head. After that size is determined, the system knows that the breast region is one-head-size down from the chin, and can map out the rest of the body from there. “It’s instinctive computing,” he said. “We use one object to measure another object.”

Iris scans

Downstairs at the CyLab is the Biometrics Center, which permits visitors in the door only after scanning their iris (the center’s director, professor Marios Savvides, also has a key.) That same kind of iris scan technology is seen by Dr. Savvides and his twenty student assistants as a viable and eventual alternative to fallible security methods, such as body scanners and behavioral analysts.

 

The problem is that