X-Ray Scanner Vision (01.14.05) - Cargo containers could be used to smuggle nuclear material into the U.S., experts say. But scientists are building a scanner that might detect nukes using part of the same process found in nuclear reactors.
Terror Family (07.23.04) - The Olympics are weeks away and organizers are scuttling to further tighten security. Researchers say terrorist groups are doing organizing of their own, using a trait found in evolution to train suicide bombers
Unsecured Radiation (09.12.03) - With the concern over dirty bombs, youd expect the security of radioactive material to be top notch. But that doesnt seem to be the case at university campuses.
The practice of fingerprinting foreigners as they enter the United States in order to stop terrorists from legally entering the country may not be as effective as the government hoped. As this ScienCentral News video explains, a just-released study says the system must better account for differences in human fingerprints.
U.S.-VISIT
The $10 billion dollar U.S.-VISIT program is the Department of Homeland Security's plan to stop terrorists from simply catching an airplane, or a boat or just driving into this country. Holders of foreign passports must place first their right and then left index fingers on a scanner that records the print. They are also photographed upon entry. The prints are then quickly compared against a government watch list. If the visitor's prints are close enough to one on the watch list, the visitor is taken aside for further identification and questioning.
But, the current system, "simply isn't satisfactory for that level of investment," says Stanford University professor Larry Wein in a study just released by the journal Proceedings of the National Academy of Sciences. He says, "It turns out that different people have different quality of their fingerprint images. Some people have quite defined grooves and valleys and other people either genetically or because they've worked hard labor most of their life have quite worn-out fingers."
He adds, "about ten percent of the people on the federal watch list have the poorest quality fingerprint images. And for those people, if they came in and gave their true fingerprints, the chances we would catch them are only about 53 percent, so barely more than a coin flip."
Wein says there are two things that should be done to fix the system. The first is switch from just two fingers to eight or ten. He says, "By doing so we could increase the chance we'd catch them to 95 percent, without increasing the chances that we would falsely identify people who are not on the watch list." But that will be expensive because the machines rolled out a little more than a year ago can only scan one finger at a time. The second, cheaper fix, says Wein, will only require rewriting, "a few lines of computer code" to change how low-quality fingerprints are assessed as possible matches. According to Wein, "The current strategy is a 'one size fits all' strategy, regardless of the quality of the fingerprint image, they use the same threshold. We're suggesting as a second software fix that you could vary the threshold levels for different image qualities essentially making it more difficult for people with poor image quality to evade the system." By doing that, he says, "we could increase the detection probability of the lowest image quality people from 53 percent as it is in the current system up to 73 percent."
Wein says there are two things that should be done to fix the system. The first is switch from just two fingers to eight or ten. He says, "By doing so we could increase the chance we'd catch them to 95 percent, without increasing the chances that we would falsely identify people who are not on the watch list." But that will be expensive because the machines rolled out a little more than a year ago can only scan one finger at a time. The second, cheaper fix, says Wein, will only require rewriting, "a few lines of computer code" to change how low-quality fingerprints are assessed as possible matches. According to Wein, "The current strategy is a 'one size fits all' strategy, regardless of the quality of the fingerprint image, they use the same threshold. We're suggesting as a second software fix that you could vary the threshold levels for different image qualities essentially making it more difficult for people with poor image quality to evade the system." By doing that, he says, "we could increase the detection probability of the lowest image quality people from 53 percent as it is in the current system up to 73 percent."
Wein's area of expertise is Operations Research, and is editor-in-chief of the main technical journal in the field. Operations Research applies complex mathematical formulas and statistics to analyzing complex problems, ranging from how to better make fast-food to operating an airline. Wein says, "I view it as developing mathematical models of real systems with an eye towards improving these systems, making decisions to better manage these systems." Prior to the attacks of Sept. 11, 2001, Wein used his knowledge to improve the efficiency of businesses, but since then he's helped the government improve security efficiency in two areas. The first was helping Washington, D.C. design a method to distribute vaccines in case of a bio-terror attack and the second was improving shipping container security.
U.S.-VISIT was introduced in late 2003 and expanded to 115 airports and 14 seaports in January, 2004. All U-S ports of entry are to have the system by the end of 2005.
delicious
digg
magnolia
reddit
newsvine
netscape
netvibes
