Slashdot Mirror
Vein Patterns to Verify Identity
JonN writes "Fujitsu Ltd. will start selling a biometric security device next month that relies on vein patterns in the hand to verify a user's identity, it said today. The palm-vein detector contains a camera that takes a picture of the palm of a user's hand. The image is then matched against a database as a means of verification. The camera works in the near-infrared range so veins present under the skin are visible, and a proprietary algorithm is used to help confirm identity. The system takes into account identifying features such as the number of veins, their position and the points at which they cross."
Biometrics sounds great, right up until the point you run into the desperate dude who is willing to take out your eyeball -- or in this case remove your hand -- just to be able to access whatever it is that is being protected by biometrics.
So who is this really good for?
Wouldn't you rather give up the memorized password rather than your eye or your hand?
But then, how does your employer look at this.
He doesn't give a shit about your body. He just wants to protect corporate assets. From his point-of-view, it is statistically less likely that he'll lose such assets were biometrics used over passwords.
Just remember that when next you go to ask for the raise, and your boss is making you authenticate to the company's grid using biometrics.
I guess more biometric sensors are always better -- but at a point, doesn't it seem excessive? I guess I'll be able to sleep easier tonight knowing that if I'm killed in my sleep and my murder spreads my bodyparts across the county, I can still be indentified by the veins in my hands. Thank God.
It sounds like such a system is subject to replay attacks. i.e., if I take a picture of your hand I can replay it to the reader. That's the beauty of smart cards: challenge-response with a random nonce means no two queries are ever identical, so no two replies are ever identical, and replay attacks are worthless.
That's the dumbest argument I've heard all evening.
The "desperate dude who is willing to take out my eyeball?" Why wouldn't he just leave it in your head and just piggyback through? Or bring you along to access that "protected" stuff?
Sure I'd rather give up a memorized password instead of an eye or hand, but again this is a question of severity. I don't believe you go from demanding a password to cutting out an eye without things other than biometrics being a critical factor.
Your employer may not give a shit about you, but most employers do. The liabilities of employees getting hurt is much of the reason that many employer-offered health plans have increases every year. I doubt that any employer will be nonchalant when one of their employees come to work with only one hand.
There's nothing wrong with an employer implementing biometrics, if it's an at-will company. It's up to the employee as to whether that proposition is acceptable.
I find it worrisome that the verification of something as personal and important as someone's identity is based on something as common and repeatable as the pattern and layout of veins.
I haven't done the research, but I doubt this is any more "repeatable" than fingerprints, or for that matter DNA.
I've met quite a few people who have nonstationary veins; that is, veins that they can move around, that twist under their fingers and stay in their new position, etc.
How will this system handle these?
It's only an insult if it's not true.
Well, I see we've already got a few people posting "zOMG my hand's gonna get chopped off".
Here's a pop quiz. How's a device that uses near-IR to see active blood vessels going to work....
...on a hand with no blood pressure, and no hot blood flowing through it? Seems to me a cut-off hand would be virtually worthless within seconds; the veins would become the same temperature as the rest of the hand, and collapse due to lack of blood pressure.
Please help metamoderate.
This time, it's the translucent map of the hand.
Problems with this idea?
1. Injury or other causes of restricted bloodlow will change the pattern. People may be wearing a watch or carring a bag which may change the net translucent image of the hand for some time.
2. No mention if this is 3-d imaging, or multiple-perspective scanning of some sort - but if it's just a 2-d single image, then another source of the 2-d image could be used as fake ID. In the case of 3-d imaging, fakes become more difficult - gummy hands are a lot less common than gummy bears. Still - there has to be a basis for pattern-recognition in the complex mess that makes up a human hand/palm, and that basis can be exploited. A rubber glove with ink on the palm, flipped inside-out may do the trick, or something similar.
3. This equipment... will it be cheap? Will it require large databases and further security for that data? How much cheaper will this be than other security methods? Cost more than most things will likely determine the impact of a biometric technology. Just having another identification scheme won't help that much, if it can only be used in already-secure or expensive scenarios.
Biometrics are a great idea, and some very cool implementations - but they always seem to involve a lot of false negatives/positives (none have solved both), and are fairly expensive relative to their unreliability. They certainly haven't been a replacement for most standard security schemes. How is this scheme different?
> Some day in the very near future there will be a way to easily duplicate fingerprints, vein prints, retina prints, or whatever.
Some day in the very near future, there will be biometric scanners that can tell the difference between real/live and fake/amputated body parts. The fact that there are not many now is mostly due to the fact that nobody wants to pay for them. People seem to think that spoofing is not an issue. But it is, or will be. As biometrics are increasingly used to protect things of value (cars, credit cards, etc), it becomes more profitable to develop spoofing techniques. That, in turn, makes it more profitable to develop better liveness detection methods. It's an arms race, really.
You guys are all overreacting -- as if this will be the end all be all of identification.
This won't be used solely except perhaps for minor barriers to entry. You don't need to worry about some guy having the same vein pattern as you, since the chance that this guy is also trying to defraud you is pretty small. A criminal might share a pattern with some other people, but how is he going to find out which people he matches without some inside access to the system?
You people worried about not reading due to various biological reasons: it may be an inconvenience, but you aren't gonna be locked out of your account. What do you do if you forget your password nowadays?
And those who say that the system is insecure and bypassable. No system is secure. At least this is probably more secure.
But you don't want to have to do this every week, for practical and security reasons.
It'd be like changing your password every week, automatically, doesn't seem like so bad of an idea. Just a pain to maintain.
What?
It didn't seem to be mentioned, but there are more details regarding this device people ought to know. The current version they are pimping is a black block a little bigger than a pack and a half of cards, and while it is true it does IR vein scanning, it also is performing hand geometry matching in software, to assist in getting around vein irregularities. The Fujitsu minions who use it seem to like it, and have gotten so good at positioning their hand over it that they no longer need the positioning guide. They are already prototyping a new version that is roughly a 1 inch cube including the USB controller.
No matter how much they work in software to prevent fakes from gaining accees (which by the way is pretty decent from the demo I saw) it is still weak to the picture attack fundamentally. The algorithms are pretty good though, since a major japanese bank is using them on ATM's, combined with a smartcard ATM card that does MatchOnCard with the calculated template from the sensor (ostensibly to prevent private information from leaking to the outside, though they still require a PIN as well to use the ATM).
Now if only the manufacturers moved to a harder to fake biometric sensor type, like ultrasonic fingerprint sensors...
My main problems with almos all biometrics identification & recognition systems for public use is that
- none of them works good enough (see below)
- if you combine multiple biometrics to raise the efficiency they will become exponentially more inconvenient and expensive, and still not being 100%
- very many biometrics can be falsified and there probably are levels where even cutting a hand isn't a big deal to get to the information; in cases when you need the hand/finger/etc. alive there's kidnapping and remember, one doesn't have to interrogate the fella, just to take him
Ok, so about efficiency. If you care to dig a bit deep and read research regarding different types of biometrics, you'll easily find quite high numbers on %. There's two things one has to constantly keep in mind:
- most if them give those high % only in specific working conditions
- if you read one biometrics works at 9x%, always think on the reverse: e.g. how many real people does that 100%-9x% mean in the real life like airports with multi-million guests a day ? even 99% goodness means 10000 from 1mil. people falsly angered and that's a lot
I am putting myself to the fullest possible use, which is all I can think that any conscious entity can ever hope to do.
The real problem here is the false negatives. Suppose I switch from typing to writing for a couple of weeks. Two weeks later, all my viens have moved back into the base of my palm and away from the little finger. It's too temperamental compared to ascii passwords :)
If I end up implementing unbreakable security somewhere , it's be proximity card (RFID) + password + biometric. This combines - what you have, what you know and what you are. Also some very good error messages if you type the password wrongQuidquid latine dictum sit, altum videtur
For 99.99999% of the applications out there, no one would even DREAM of going to these lengths.
For the other 0.00001% (read military secrets) of the applications out there, there is likely to be two or three other authentication processes out there, one of which involves a person pysically giving you access.
See my journal for slashdot ID's by year. Mine created in 2005. http://slashdot.org/journal/289875/slashdot-ids-by-year
But looks really cool in movies.
Anything that can be imaged can be reproduced to the accuracy of the imager. Hence, biometric security is like a social security number: it might be unique to you, but you can't change it ever* and if someone gets a hold of it, you're screwed.
*I am aware that in extreme situations you can change your SSN. afaik, This capability was designed to address that point, however the address space of SSNs is not that sparse and the cost of changing the number is too high. (in both time and money)
The only way to change your biometric data would involve some pretty severe scarring.
Can you be Even More Awesome?!