Stories
Slash Boxes
Comments

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

Vein Patterns Could Replace Fingerprints

Posted by samzenpus on Thu Nov 13, 2008 04:21 AM
from the no-two-alike dept.
Death Metal writes "Companies in Europe have begun to roll out an advanced biometric system from Japan that identifies people from the unique patterns of veins inside their fingers. Finger vein authentication, introduced widely by Japanese banks in the last two years, is claimed to be the fastest and most secure biometric method. Developed by Hitachi, it verifies a person's identity based on the lattice work of minute blood vessels under the skin."
+ -
story

Related Stories

This discussion has been archived.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

    it's big and blue.

  • Makes you wonder what else can be discerned from the pattern of blood vessels and other scan information.

    Can't let all that valuable information go to waste, can we?

  • by mlts (1038732) * on Thursday November 13 2008, @04:48AM (#25744569)

    Maybe its me being pedantic, but I consider biometrics something that is intended to replace typing in a username, as opposed to being both pairs of the username/password combo. Ideally, one would have biometrics to ID which user is wanting access, then have a contactless smart card and/or a PIN for the "password" part that confirms the user is whom he or she said they are.

    • by cjfs (1253208) on Thursday November 13 2008, @04:59AM (#25744609) Homepage Journal

      Agreed. Single-factor authentication based on something that's not reissuable is a recipe for failure.

      Eventually people will run out of non-compromised fingers ;-)

      • Agreed. Single-factor authentication based on something that's not reissuable is a recipe for failure.

        Very true, and very bad. On top of that, add a horribly broken authentication protocol:

        You send your username to the authenticating party in the clear, and they verify that it matches their stored copy of your username.

        Hello... ? What's the password, here? I can't think of any way to copy your fingerprint off that laptop I just stole from you. Nor do I ever get the idea to produce a workable replica of your iris from the hi-res photo I have in my database.

        On the web, where you send your site-specific pa

    • We used to have something like that called Ugly People when I was a kid.
    • Re: (Score:3, Interesting)

      Your right, should be a two-step process.
      The summary uses two terms, identification & authentication, as if they were interchangeable. They are not.

      Identification is the process by which the identity of a user is established, and authentication is the process by which a service confirms the claim of a user to use a specific identity by the use of credentials (usually a password or a certificate).

      So the biometrics would identify you, not authentify.

    • I consider biometrics something that is intended to replace typing in a username

      And wisely so. Biometric data is an identifier--it's something with a one-to-one mapping to an identity (here: a pile of cells). Other common identifiers are SSNs, usernames, user IDs, RSA public keys and sha1 hashes [the one-to-one-ness works well in practice for sha1, but of course not in theory].

      Identifiers are not authenticators. A good authenticator for any given identifier requires that only the identified thing can produce the authenticator; except in one-time schemes, performing the authentication should not allow anyone else to authenticate as you later on. It also requires that they one you're trying to prove something to can verify what you're claiming.

      A good authenticator for a public key is a signature on a random string. [make sure the one validating you knows how the signature looks before you send it; use a commitment scheme].

      A bad way to authenticate is by sending a copy of the private key [or for sha1 hashes, the string that hashes to the given hash].

      Biometric authentication "works" by having the identifier be the authenticator, and the authentication protocol works by sending a copy of the authenticator:

      You put your iris in front of the scanner and it does a "SELECT permissions FROM users WHERE iris = %s" [without the horrible SQL injection possibilities, of course]. What's to stop those who look up your iris from creating a replica? If you work by fingerprints, I send my goons to follow you around. When you open or close a door, they take your print and produce a rubber replica.

      An analogy would be that you learn a word that only you can pronounce, and the authentication works by you saying the word aloud, such that everyone in your vicinity can hear it. "Only you can pronounce", I don't buy that.

    • Re: (Score:3, Informative)

      I don't know about this particular method, but biometric measurements in general are not perfectly repeatable, so they need to use fuzzy algorithms, which raise the probability of collisions. So they are more like a hashed password than a perfectly unique user ID.
    • by DrYak (748999) on Thursday November 13 2008, @10:00AM (#25746695) Homepage

      Plus, as an MD, I have quite some suspicion about the stability of some biometric methods over time or over pathologies.

      Take today's method :
      - it relies on vein patterns.
      The main problem I see is that veins are biomechanically elastic, in order to be able to comply with varying amount of blood. It works as a "blood pool".
      Depending on pathologies, the shape of the veinous network can change dramatically.

      (same goes for retina. I mean looking at the change induced is the way to assess the progress of some disease like diabetes or hypertension).

      Fingerprint worked so-so because the relatively stable : as long as the deeper structures aren't destroyed, the skin regrows with the same prints, no matter what.
      Fucking up fingerprints require deep mutilation of fingers. These kind of accident can happen is heavy industrial workers, but its not something the average laptop wielding geek is very likely to experience. Thus fingerprints are good enough.

      Whereas, the current trend of blood-related biometric systems are affect by pathologie (I've mentionned hypertension and diabetes) which are much more frequent, specially among the sedentary people: typically the users of such systems.

      Thus, I have real doubts about the long term feasibility of such measures.

      • Messing up a fingerprint requires nothing more than a sharp object and several horizontal lacerations in some cases... I know this from personal experience. On of my fingerprints was permanently altered enough that the whorls were distorted when I got a simple cut that became infected. The regrown print now has a section across it stretched to the side, distorting the shapes enough that most systems don't recognize it as the same fingerprint anymore.

        Of course, there is still enough that a human can identify it, but the limited data sets used in most biometrics can't find sufficient matching markers.

        In the case of another finger, I also have vertical wrinkles that come from aging, so now that fingerprint is segregated like looking at it through blinds.

        There are many everyday events that can cause enough change in fingerprints to mess up most biometric readers. These range from short term events like having a cut or blister, to permanent changes like slicing a fingertip off and the doctor not lining it back up perfectly.

        There is nothing about the human body that is immune to change. It is that elastic ability to adapt that has made homo sapiens a viable species.

  • by irexe (567524) on Thursday November 13 2008, @04:58AM (#25744603)
    Until someone figures out how to revoke and replace biometric properties in case of fraud, I don't see why we should even be considering them as a serious replacement for good old passports.
    • by phoenix321 (734987) * on Thursday November 13 2008, @05:16AM (#25744695)

      It would be very necessary to mandate a "duress PIN" or password for every authentication point. A silent alarm whenever someone is forced to enter credentials against their will.

      This should be mandatory for all authentication systems anyway, it would certainly hinder these ridiculous one-day kidnappings and ATM muggings.

      • Re: (Score:3, Interesting)

        And how would a duress pin help anything?
        As if the cops could jump onto the scene during the short time that an ATM transaction takes...

        If the bad guys stand next to you, pointing a gun to your head while you make the transaction then the ATM camera will capture that anyways and provide good evidence later on.
        But if they don't (which, I guess is more likely) then entering a duress pin changes exactly nothing. Sure, the bank now knows that you may be in trouble - but what can they do, hand out marked money?

        I

        • Re: (Score:3, Insightful)

          And how would a duress pin help anything? As if the cops could jump onto the scene during the short time that an ATM transaction takes...

          Seriously? In the case that a duress code is entered, the police have a lot of information to work with. As opposed to someone reporting the crime possibly hours after is happened. 1. The cops are informed that a kidnapping is in progress right now and have the exact location of the kidnappers. 2. They know the person who has been kidnapped by the bank account that is being accessed. 3. They have the images from the ATM camera, which may indicate how many kidnappers there are, how they are dressed, what

    • Re: (Score:3, Insightful)

      One key element of that method is that fraud is harder to perform than with other method of biometric identification methods. You leave fingerprints and DNA samples all the time and they are easy to copy or displace, and yet they can and are used as strong evidences in criminal cases. At least, with vein patterns, no one can copy yours from an indirect transfer on a regular surface of from a photo of you.

  • by (Score.5, Interestin (865513) on Thursday November 13 2008, @05:05AM (#25744637)

    An evaluation by the National Physical Laboratory in the UK found vein patterns to be the least reliable biometric they'd ever encountered, worse even than face recognition which became notorious for its zero-percent hit rate in several public trials (OK, so you can't get worse than zero percent, but in carefully controlled lab trials face recognition did get a non-zero score).

    Looks like another great example of biometric vendor marketing at work. "Buy our stuff, it's gooder than anyone else's!".

    • Re: (Score:3, Insightful)

      I don't see the contradiction: I would consider the least reliable metric the most secure.
      Or are we talking about the security of the bank?
    • by smoker2 (750216) on Thursday November 13 2008, @01:06PM (#25749477) Homepage Journal
      In my capacity as truck driver, I have had occasion to visit Felixstowe container terminal. They have been trying to get a similar system going for years. I have a photo card that contains the data and I have to place my hand on a pad up against metal posts. This system has never worked reliably, and so far other than when I went through the initial process, I have never had to use it. The terminals are always out of order. So we just wave the card instead.
  • Bonus news (Score:3, Interesting)

    by Artifex (18308) on Thursday November 13 2008, @05:06AM (#25744641) Journal

    It's less likely your fingers will get hacked off and taken by criminals trying to get past scanners, if this is used. Although I suspect criminals will find a way to flash-freeze fingers, seal the ends, and then warm up in water before using in the same situations where they could get away with severed fingers for fingerprints (remote access, etc.)

  • by hcdejong (561314) <acme@xm[ ]t.nl ['sne' in gap]> on Thursday November 13 2008, @05:08AM (#25744649)

    The gruesome possibility that criminals may hack off a finger has already been discounted by Hitachi's scientists. Asked if authentication could be "forged" with a severed finger, the company says: "As blood would flow out of a disconnected finger, authentication would no longer be possible."

    So you'd need a contraption that feeds blood through the finger. It's an extra obstacle, but if you're desperate/psychopatic enough to sever someone's finger, rigging a blood supply is no big obstacle.

  • So, is there actually a working, reliable, automatic fingerprint-reading system that can be used as legal evidence? I am not aware of one, but I am not aware of a whole lot of stuff. (like where's my coffee?)

    Cragen

  • Well, at least this method offers less excuse to post gross pictures on Idle [slashdot.org]. So I'm all for it. Extra points for being able to give the bank machine the middle finger (yes, I've actually read the article).
  • They recently introduced the palm scan to ID people walking in and out of their tests (GMAT etc). I still haven't figured out why. If nothing else it's an interesting way to get strange diseases from sick people who sneeze politely.
  • by Viol8 (599362) on Thursday November 13 2008, @06:23AM (#25745017)

    ...getting sick of the endless ways to identify and tag individuals that have appeared recently? Fingerprints, iris scans, voice recognition, face regonition, smell (!) , walking gait, now vein patterns. How long before we're all just barcoded with a unique id??

    I'm sure some people will say I'm just being paranoid but with the advancement of AI image processing it won't be long before we can be identified no matter where we are , what the time is , or what we're doing. Yes , the governments all roll out the "terrorism" line whenever questioned about this but we've all seen how its been abused already.

    So whats next - infra red heat pattern signatures of individuals? Chemical piss analysis in public toilets?

    • by squoozer (730327) on Thursday November 13 2008, @06:54AM (#25745149) Homepage

      We sort of already do carry around a barcode - in our DNA. While we aren't even close to being able to process it fast enough to make it viable at the moment I could easily imagine we will be able to in the future. Welcome to the world of Gattaca only we won't be able to get round the checks as easily as he does in the film.

  • I forget their name, but I actually used a working prototype that used this exact method of biometric identification.

  • by floydman (179924) <floydman@gmail.com> on Thursday November 13 2008, @06:53AM (#25745147)

    For the plain simple fact that they leave traces behind. Police work, you know!

  • Identification and surveillance technology is advancing very fast. There was a story a few weeks back about how keys can be copied from photographs, and I expect eventually Minority Report style eye scanners that work from a distance will become available. Maybe even fingerprints from a photograph.

    AI is improving quickly too, and I expect eventually a computer will be able to take feeds from various cameras/scanners/RFID and use it to track a person automatically. At that point the UK government will want t

  • Not New (Score:3, Informative)

    by Thnurg (457568) on Thursday November 13 2008, @09:52AM (#25746583) Homepage

    This is not new. Vein pattern recognition on the back of the hand was developed years ago. So long ago in fact that the computer part of it was a BBC Micro. [blogspot.com]

  • by Assmasher (456699) on Thursday November 13 2008, @10:32AM (#25747109) Journal

    ...been doing this specifically for security biometrics for years. Perhaps the news would be that it will become more pervasive, but the same problems that prevented it from taking off in the past apply now as well - you have to network the device in order to validate the user's pattern (most of them actually create a sort of hash code actually.)

    • by Roland Piquepaille (780675) on Thursday November 13 2008, @04:34AM (#25744493)

      Funny you should say that, my first thought when I read "Finger vein authentication, introduced widely by Japanese banks in the last two years" was that it's got to be a bitch to withdraw cash from an ATM if you're a Yakusa...

      • Re: (Score:3, Interesting)

        I am under the assumption that it would rather difficult to get close enough (contact) to someone and use a special light and scanner/sensors to obtain vein patterns without a person knowing... except if maybe they were asleep. This isn't a photograph, it is a contact scan that requires multiple infrared light sources.

        From a security standpoint, even if you did obtain someone's scan, then how exactly would you impersonate a fake vein pattern in your arm to trick a scanner?

        I do want to point out that I woul