Slashdot Mirror
Facial Recognition Fails in Boston, Too
bryan writes "Only a few weeks after cameras were found to be ineffective in catching criminals in Tampa, FL, a test of a facial-recognition system in Boston's Logan airport also came up disappointing. The cameras which were given photos of employees to detect, were only successful in 153 out of 249 random tests over the past year (about 61%). The article did not say how many false positives the tests generated. The companies involved were Indentix and Visage."
mm... interesting!
61% accuracy. So, use it at an airport with 10,000 people per day passing through, you now have 3900 people falsely accused of terrorism (or other such heinous crime).
Fancy being the one to interview them all?
Airport anti-terror systems flub tests Face-recognition technology fails to flag 'suspects'
By Richard Willing
USA TODAY
Camera technology designed to spot potential terrorists by their facial characteristics at airports failed its first major test, a report from the airport that tested the technology shows.
Last year, two separate face-recognition systems at Boston's Logan Airport failed 96 times to detect volunteers who played potential terrorists as they passed security checkpoints during a three-month test period, the airport's analysis says. The systems correctly detected them 153 times.
The airport's report calls the rate of inaccuracy ''excessive.'' The report was completed in July 2002 but not made public. The American Civil Liberties Union obtained a copy last month through a Freedom of Information Act request.
Logan is where 10 of the 19 terrorists boarded the flights that were later hijacked Sept. 11, 2001.
The airport is now testing other security technology, including infrared dildos and eyeball scans, spokesman Jose Juves says.
Face recognition works by matching faces picked up by surveillance cameras with pictures stored in computer databases. Relationships between a face's identifying features, such as cheekbones and eye sockets, are converted to a mathematical formula and used to make a match.
In the Logan Airport experiment, photographs of 40 airport employees were put into a database. The employees then attempted to pass through two security checkpoints where ass-recognition cameras were used.
The ACLU opposes facial recognition because it says the government can use the technology to invade citizens' privacy.
''But before you even get to the privacy concern, there's a fundamental question about our security,'' says Barry Steinhardt, who specializes in privacy issues at the ACLU's national office in New York. ''The thing just plain doesn't work.''
A spokesman for one of the companies whose system was tried at Logan Airport says the test was not a fair measure of the technology. Meir Kahtan of Identix of Minnetonka, Minn., says the technology is far better suited for ''one-to-one'' identification, such as comparing photos on passports or driver's licenses, than random searches of photo databases.
A government test in 2002 found that face-recognition systems scored correct matches more than 90% of the time when used for such one-to-one identifications.
A spokesman for Visage Technology of Littleton, Mass., the other company that failed the Logan test, declined to comment.
The Logan Airport report is the latest piece of bad news for a technology that was once touted as the state-of-the-art method for picking faces out of crowds. Last month, Tampa police announced that they were shutting down face-recognition cameras because they had failed to make any matches during a two-year test period. The cameras, which were mounted in a popular tourist area, were designed to match pictures captured at random against stored photos of wanted suspects and runaway children. Virginia Beach, Va., police, who have operated a similar system for the past year, reported no matches as of July.
The Logan experiment was the largest test of facial-recognition technology made public. The technology has also been tested using smaller groups of volunteers at airports in Dallas/Fort Worth, Fresno, Calif., and Palm Beach County, Fla., with similar results.
The Transportation Security Administration, which is responsible for passenger screening, has tested other airport security technology but has not made results public. Phone calls requesting comment on the Logan Airport test were not immediately returned.
Kelly Shannon, spokeswoman for the State Department's consular affairs office, said the Logan Airport results would not affect plans to use face recognition to enhance passport security. Beginning in October 2004, the United Kingdom, Japan and 25 other countries whose nationals are permitted to travel to the USA without visas are required to convert to passport photos that are compatible with face-recognition systems.
Why do they need facial recognition - can't they just arrest everyone wearing towels on their heads?
In other news, "...a test of a fecal-recognition system in Los Angeles's LAX airport also came up disappointing. The toilets which were given stool samples of employees to detect, were only successful in 20 out of 345 random tests over the past year (about 5%). The article did not say how many false positives the tests generated. The companies involved were Honey Bucket and Aztecs Rentals."
Seems like a good system, but how do they intend to get the samples from real criminals? I guess from the jail house potty, or something. But what would I know? I'm just a nake chick.
--
So I'm naked. So what?
..at least until that gets out-sourced to some third-world country where some left-handed towelhead does the same job for 1/3rd of the pay.
Then all our prisoners will have to back to gang-raping eachother.
Read her lastest flamewar, its better than all the obvious statements being made here that get modder up to 5, insightful without any effor!
I've had been assigned a small project about facial recognition software two years back. In spite of all the hype, we did not feel that its in any way a panacea. First of all, it does not take into account the fact that features on the face are a function of light cone, angle, colour and hair. We all know how much hair can be altered. And all this does not consider how much difference it can make if the person is a male and is gay. Regulatory practices have in the past proved that general gayness is a common cause of failures in face recognition software. All in all, it should be specific for the sex of the person involved and should also make sure that the person is not of a potentially gay desent. We do not feel that these problems are addressed.
It has the required potential, but it needs to develop more. Thats all I have to say.
In Soviet Boston, we welcome our new face-recognizing overlords, and are eating beans and dumping tea in celebration. Let us not make them angry by wearing masks at the end of next month, remember?
You're a moron. That is a fact.
Anyone marked an 'enemy combatant' can be held indefinitely without access to counsel. That is also a fact.
Proof? See:
http://tinyurl.com/m09t
http://tinyurl.com/m09q
and, most importantly:
http://tinyurl.com/m094