Stories
Slash Boxes
Comments

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

Public Iris Scanning Device In the Works

Posted by kdawson on Wed Feb 07, 2007 09:04 AM
from the road-you're-on-john-anderton dept.
Nonfinity writes "A public iris scanning device has been proposed in a patent application from Sarnoff Labs in New Jersey. The device is able to scan the iris of the eye without the knowledge or consent of the person being scanned. The device uses multiple cameras, captures multiple images, and then selects the best image to process."
+ -
story
This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More
Loading... please wait.
  • by Anonymous Coward on Wednesday February 07 2007, @09:07AM (#17919534)
    Damnit, someone watched Minority Report and went "Heeeeey, good idea....GET ME R&D"
    • Re: (Score:3, Interesting)

      I know you're joking, but on a semi-serious note...doesn't that movie constitute 'prior work' or 'prior invention'?

      The patent system, as it's defined, says that patentable ideas must not be a logical extension of existing ideas or an idea already created by somebody else. I skimmed both links and I can't find (maybe I missed it?) any mention of a the date related to when this company claims first provable conception of the idea. Unless they built something years ago, this isn't going to hold water.

      W
    • I foresee a new market developing for iris-concealing contact lenses.
  • Put on... (Score:4, Funny)

    by FredDC (1048502) on Wednesday February 07 2007, @09:09AM (#17919552)
    Put on your tin foil hat... And sun glasses!
  • Priorities? (Score:5, Insightful)

    by blowdart (31458) on Wednesday February 07 2007, @09:09AM (#17919554) Homepage
    Interesting to note that the article focusses on the less sinister uses for this, customised advertising, whilst bypassing any mention of privacy aside from a nod to saying it could take place "without the knowledge or participation of the subject". So whose money will talk fastest, advertisers or Homeland Security?
  • Won't Work (Score:5, Interesting)

    by giafly (926567) on Wednesday February 07 2007, @09:12AM (#17919584)

    The problem, says Davies, is the limited accuracy of biometric systems combined with the sheer number of people to be identified. The most optimistic claims for iris recognition systems are around 99 per cent accuracy - so for every 100 scans, there will be at least one false match.

    This is acceptable for relatively small databases, but the one being proposed will have some 60 million records. This will mean that each person's scan will match 600,000 records in the database, making it impossible to stop someone claiming multiple identities. - new scientist [newscientist.com]
    Please can someone design one of those standard forms for these bogus ID schemes - like the one with all the reasons why anti-spam technologies won't work.
    • I've never seen such pessimistic claims for iris recognition. With a false accept rate of 1/1000 to 1/10000, you can achieve a false accept rate of pretty much zero. I respect Simon Davies, but I'm not sure he has his facts right here.
    • Re:Won't Work (Score:4, Insightful)

      by M_Hulot (859406) on Wednesday February 07 2007, @09:35AM (#17919786)
      You're quite right that it won't work. The 99% accuracy figure that you quote is very high, compared to fielded system. The UK government seems to have put it's scheme on hold after it "failed half its assessments." http://www.theregister.co.uk/2007/01/11/project_ir is_evaluation_report/ [theregister.co.uk] Note that these tests were on actively cooperating participants. The success rate for those not cooperating has to be very low.

      BTW the Live Science article suggests that: "Good quality scans result in a "false match" less than one time per one hundred billion". This estimate seems to be off by a factor of between 1 and 10 billion. Check out other articles by the same journalist: "New Study finds Sun only 491 feet from Earth".

  • by zyl0x (987342) on Wednesday February 07 2007, @09:12AM (#17919586)
    In all seriousness, I would've thought someone in London would come up with this idea first.
      • What's the point of changing unreliable technology with new unreliable technology?
        Can we at least get ONE thread that doesn't deal with "why upgrade to Vista"?
  • by Neuropol (665537) * on Wednesday February 07 2007, @09:14AM (#17919606) Homepage
    Do we know that repeated retina scanning is healthy for our eyes?
    • Re: (Score:3, Informative)

      This device is not going to scan retinas, it does iris recognition [wikipedia.org]. And no, it doesn't really do anything to your eye besides taking a hi-res photo of it.
  • by dpbsmith (263124) on Wednesday February 07 2007, @09:19AM (#17919632) Homepage
    Contact lenses that alter eye color are already in popular, widespread use.

    How hard would it be to construct a contact lens with a unique, fake, computer-generated iris image (no idea how you'd do that, but "fractals" sounds like a good buzzword to insert here)? Sound like it would be a lot easier than fake fingerprints.

    In a situation where you knew you were being scanned, the officials might say "I see you're wearing contacts, remove them please," but I don't quite see an airport saying "no contact lenses allowed in this airport..." particular if the idea is that the scanning is supposed to be surreptitious.

    • by dpbsmith (263124) on Wednesday February 07 2007, @09:29AM (#17919726) Homepage
      The article says "Good quality scans result in a 'false match' less than one time per one hundred billion."

      It also says "the newly proposed system is that it allows iris scans to be taken without the knowledge or participation of the subject."

      What it does not say is that "the newly proposed system allows good quality scans, with a 'false match' of less than one time per one hundred billion, to be taken without the knowledge or participation of the subject." I fancy readers are supposed to infer that conclusion, which does not follow from the premises.

      I'll bet the system has the usual impressive-sounding "99.9%" accuracy or something in that ballpark... like all those facial-recognition systems. Meaning a false positive rate of one in a thousand. Meaning that if one in a million airport visitors is a known terrorist with an iris scan in the database, then 999 out of every thousand people, yanked out of the concourse by polite but firm security officials, will be Lutheran grandmothers from Davenport, Iowa travelling to visit their children in St. Paul.

      And the officials will be unable to give any coherent explanation, since the system is supposed to be surreptitions.
    • Re: (Score:3, Informative)

      Already have them, although the iris isn't terribly natural.

      http://www.9mmsfx.com/lenses.html [9mmsfx.com]

  • by Shivetya (243324) <shivetya@archo[ ].com ['non' in gap]> on Wednesday February 07 2007, @09:22AM (#17919658) Homepage
    specifically about implementing something this.

    Identifying who you scanned. sure you can scan an iris without their knowledge but unless you have the pattern stored how will you know who it is? Perhaps do it at a register and match it to the card/id used? That would be underhanded to say the least.

    Storage, how much space per pattern? What is the speed of comparison to a large database? Something that is quick enough to focus ads (for the minority report fans) would require serious processing power.

    I could see it in small settings, say a business who needs a less instrusive means of security. Scan all your employees and only let them in, if accompanied by those who cannot be matched then don't admit to sensitive areas. However in the general public setting, costs for equipment to store millions of scans and process them fast enough to be meaningful is still aways off.
  • Jab (Score:3, Funny)

    by Talisman (39902) on Wednesday February 07 2007, @09:30AM (#17919736) Homepage
    I don't know about the rest of you, but iris scanners scare the crap out of me. Every time I look into the peephole to get scanned, I'm relatively certain a large needle will shoot out from behind the glass and stab me in the eye.
  • by Cicero382 (913621) <clancyj@tEINSTEI ... minus physicist> on Wednesday February 07 2007, @09:34AM (#17919774)
    ...is not so much that this is possible, but that the inventors seem to feel sure that there's a market in this AND that there won't be any serious objection to stop it. A bit like the proliferation of "security" (read "unadulterated snooping") cameras in London.

    Actually, thinking about it, what *really* worries me is that people *won't* object to it. Not really.

    Ah! Brave new world... etc.
  • by waif69 (322360) on Wednesday February 07 2007, @09:50AM (#17919952) Journal
    The system had a problem with people who blinked too much. I had to sit in front of the camera and remain still and it took a picture of my eye a few times before it got a good enough image. Out of 5 people who participated, all but one had to have multiple pictures taken.

    I just can't see this system being used with cameras that randomly take pictures from varying distances and work, unless the cameras and software improved quite a bit in the past two years.
  • IED (Score:3, Interesting)

    by SnackmasterMusic (1050984) on Wednesday February 07 2007, @10:07AM (#17920128)
    Why, all the better to remote detonate you with, Granny!

    Which terrorist group will detonate our beloved freedom fighters with this first?

    "and when I gave them cell phones, they could not get enough...

    generating the database is simple, just use the network of driver's license ID cameras.

    the only good news is the economics of technology mean this will be first used by high-value targets against other high-value targets. Think large-scale corporate wars vs. vengeful government agencies...with the rest of us as collateral damage.

    and- which foreign state will get access to our database first?

    on the other hand, think of how many more dead soldiers we will be able to recognize on the battle field! yay!

  • by Panaqqa (927615) * on Wednesday February 07 2007, @10:23AM (#17920308) Homepage
    Unless the system takes advantage of people that are in close proximity to the camera to get its pictures. Think about the resolution required otherwise. Let's say we have a picture that is 2,048 x 1,536 pixels... now, can you imagine a person's irises taking up more than 1% of the width of the picture, unless it were a rather close "headshot" type pose? Now, take a look at some closeup shots of human irises [drgaelriverz.com]. How much information do you think you'll get from 20 x 15 pixels?

    Now, instead of 3 megapixels, think 12. That's still only 40 x 30 pixels. Not enough.

    I'll worry when 100 megapixels becomes commonly available. (Yes I know the Navy has a 111 megapixel CCD).
    • Re: (Score:3, Insightful)

      now, can you imagine a person's irises taking up more than 1% of the width of the picture, unless it were a rather close "headshot" type pose

      Ah, we're all safe until someone invents robotically aimed telephoto cameras.

      How hard is that?
    • The right to privacy may only be implied and not specifically granted in our constitution, but it is still widely held.
      It's not about privacy. It's about human dignity.

      The constant monitoring, surveillance, identification, numbering and tagging of people in our society is an affront to human dignity. It's an affront not only to those being numbered and tagged, though they are the ones most offended, it's also a stain on the dignity of any state that permits it. Anyone who disagrees should ask people who have been tagged, with a barcode.

      But the interesting fact is, human dignity is not a universally recognised right. We've got rights to our property, lives and liberty, but not in most cases to our dignity. This is only something that has recently been awknowladged.

      The word "dignity" dows not even appear in the US constitution(enacted 1787). US citizens do not have a constitutional right to it. The Irish constitution(enacted 1937) does mention in the preamble that it is being adopted in part "...so that the dignity and freedom of the individual may be assured". But this is only in the preamble.

      Interestingly, the constitution of South Africa (enacted 1996), explicitly and unabiguously guarantees a right to dignity in Chapter 2: Section 10:

      10. Human dignity

      Everyone has inherent dignity and the right to have their dignity respected and protected.
      I guess decades of having their dignity denied to them taught South Africans that this right doesn't really go without saying. This is one ammendment I would dearly love to see in my country's constitution. (Actually the SA constitution also guarantees the right to privacy [info.gov.za] and even the right to private communications. It's an extremely progressive document which unfortunately hasn't influenced older constitutions in the way that it should.)

      Privacy in public is obviously a fallacy. But we should at least not have to suffer affronts to our dignity by being scanned and checked at every turn, or have our clothing seen through at every security checkpoint. Laws forcing Jews to wear stars or Muslims to wear crescents would probably still be constitutional in a lot of countries. A dignity ammendment would make what we know is wrong explicitly wrong. Humans aren't like animals. We have more needs than simply life, liberty and property. Dignity is one of those other needs.