Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

3D Face Imaging in 40 Milliseconds 170

Roland Piquepaille writes "Computer scientists at Sheffield Hallam University, UK, have developed a new face recognition software which can produce an exact 3D image of a face within 40 milliseconds. A pattern of light is projected on your face, creating a 2D image, from which an accurate 3D representation is generated. This technology should speed airport check-ins, but it could also be used in banks or for checking ID cards as it allows full identification in less than one second."
This discussion has been archived. No new comments can be posted.

3D Face Imaging in 40 Milliseconds

Comments Filter:
  • database? (Score:2, Interesting)

    what about the time it takes the image to be looked up in the database? i'm sure that would take it more time to verify...
    • Re:database? (Score:3, Insightful)

      switchfutguy wrote:

      what about the time it takes the image to be looked up in the database? i'm sure that would take it more time to verify...

      I think the article is implying that the real utility will be in matching the physical face with the biometric data stored on the identity card. Whatever advances are made there are independent of the verification of the card to the government registry.

      1. Create biometric data out of face holding attached to person with card.
      2. Verify face to biometric data on ID card
      3. Ve
      • "How are they going to deal with people who gain or loose a lot of weight since the biometrics were taken"

        Surely they could update the card each time it was verified. This way the changes would be extremely small and it should still pass.

        There would still be the problem if someone didn't use their card in a long time, and gained/lost a lot of weight (for example), but I am assuming that the ID cards would be used for almost everything eventually (payments, keycards, etc), so the chance of not having y
      • Re:database? (Score:3, Insightful)

        How are they going to deal with people who gain or loose a lot of weight since the biometrics were taken.

        If you can recognize your brother after he has gained weight, I'm sure the biometrics can recognize him, too. This is why the algorithms are programmed to have a margin of error. Like "it's 80% probable that this picture belongs to the person in question".
        • Re:database? (Score:4, Insightful)

          by swillden ( 191260 ) <shawn-ds@willden.org> on Saturday March 25, 2006 @11:39PM (#14996243) Homepage Journal

          If you can recognize your brother after he has gained weight, I'm sure the biometrics can recognize him, too.

          Sure, as long as the matching algorithm is as sophisticated and accurate as the one in your brain, the one developed over millions of years of evolution of a social animal, and then fine-tuned by years and decades of personal feedback training.

          The fact that you can recognize your brother after he has gained weight is evidence that it's possible to perform such recognition. It, however, in no way means that any matching algrothm we can create will match the feat.

          In this particular case (weight gain), most facial recognition matchers probably won't be fooled, as long as the gain isn't too great. Not because the matching algorithms are anywhere near as good as the human brain, but because the matchers focus primarily on bone structure, which doesn't change with weight -- unless the change is so great that it hides previously visible structure under a smooth, round layer of fat.

          In general, though, facial recognition technology is lousy. It's improving, but it's not remotely as good as what people can do.

        • If you can recognize your brother after he has gained weight, I'm sure the biometrics can recognize him, too.

          Uh, yeah, because facial recognition software is so accurate.

          I think you've watched a few too many hollywood movies.
          • I think you've watched a few too many hollywood movies.

            Or studied Neural Networks in college ;-)
            • Neural Networks should be the first indication to you that, frankly, machine learning algorithms are often only as good as their input. Training an network to recognize a face has no specific guarantee on which features it recognizes in the first place, as the nature of such an algorithm leaves it up to the algorithm to decide in the first place. Believe me; if neural networking were truly the silver bullet of getting around facial biometric variation, it would have solved the issue years ago.
              • Neural Networks should be the first indication to you that, frankly, machine learning algorithms are often only as good as their input.

                Exactly, this is why having a 3D representation of a face is going to help a lot.
          • Re:database? (Score:3, Informative)

            "Uh, yeah, because facial recognition software is so accurate."

            Actually, I've worked on 3D facial recognition algorithms and they can be incredibly accurate. You are correct that 2D facial recognition algos have limited accuracy, but that's largely because 2D cameras merely measure the amount of light reflected into each pixel, so any real-world dimensional measurement is an interpretation of feature locations and extrapolated into 3D. Perspective views is big problem as are lighting conditions, clutter

      • How are they going to deal with people who gain or loose a lot of weight

        Facial hair? I doubt this light can penetrate hair.

      • I'm still skeptical of the actual technique.

        Good idea

        How are they going to deal with people who gain or loose a lot of weight since the biometrics were taken

        Or grow a beard, or get cosmetic surgery, or have a face that swells due to medication, etc, etc.

        As someone who works on validation of computer target recognition, I recommend being very skeptical. Error rates will be very high.

        • Re:database? (Score:4, Insightful)

          by Dashing Leech ( 688077 ) on Sunday March 26, 2006 @10:07AM (#14997711)
          "As someone who works on validation of computer target recognition, I recommend being very skeptical."

          As someone who has written, studied, tested, validated, and worked with 3D target recognition and facial recognition algorithms, I agree to a point. I only agree because it's always good to be skeptical of these things until the (unbiased) evidence has been presented.

          That being said, the jump from 2D to 3D recognition is nothing but revolutionary in terms of recognition and verification accuracy. Beards, cosmetic surgery, and even swelling are not major impediments. They do reduce the match, but less than you'd think. There are areas of the face that don't change much even in such cases. 3D face recognition is more about measuring the shape of the skull. The skin generally follows the shape of the skull except where there are heavy fat deposits, or of course if you grow a beard. Unless you're changes were to the extent of a normal face to something like the Elephant Man or Eric Stoltz in "Mask" (no, not the Jim Carry movie), it's still going to do quite well.

          And even then, the application here is verification. If the changes were enough to make it fail, that would just mean you'd have to go through either a secondary manual security check and/or update the biometrics. It's not like you could, say, pretend to be Tom Cruise and say your face changed. If you don't look like Tom Cruise with a beard or fatter face, and have the same shaped skull, and have the other security info to back you up, you're going to get your biometrics updated.

          As a general comment (not specific to the parent post), I know /. likes to pick apart these things, but /. readers are not of so superior intelligence that they can come up with flaws in 2 minutes that experts in the area of research have not addressed, studied, or solved in years of research.

    • About the time it takes to get the info with the account#/SS#. Then the face would be matched against the face already in records.
    • what about the time it takes the image to be looked up in the database? i'm sure that would take it more time to verify...

      That's less important to them than to record images of everyone who flies. Gather the data first, worry about mining that data to find a patsy^Wsuspect later.

      Consider this scanner at the gates of a sporting event. You have all these people de^H^Hcontained for long enough to run them all through as many databases you want. Its doubtful their faces will change greatly during the course
    • Indeed. I work at a VsFX house and we have face models that have billions of vertices and take up over a gig on disk. How are they intending to search through a space like that with error tolerances in a timely manner?
    • I don't see the government trying to make some ultra-massive computing system linked globally. I see the main use for this would be municipal governments.. perhaps the state. How about a state has entries for all persons with a warrant for arrest? The municipal government has entries for all violent felons and drug dealers. The FBI has their most wanted list, as well as suspects currently under investigation. The US Marshalls Office has all wanted felons. None of this is a bad thing in any way. I can't imag
  • Hrm. (Score:5, Interesting)

    by Limburgher ( 523006 ) on Saturday March 25, 2006 @09:16PM (#14995791) Homepage Journal
    Impressive, but what if I shave?
    • well the same can be done easily in some other part of the EM spectrum in which the facial hair is "invisible" wether you have it or not
      • Good point. What would be really nice would be to use something where the machine only sees the skull. Even harder to fake, but there's always the possibility for medical issues. Not only would a technology like repeated x-rays casue health issues, but using some sort of intrusive body scanning could be interpreted as a HIPAA violation, unless you pre-consented in some legally binding way.
    • Re:Hrm. (Score:2, Interesting)

      by Watson Ladd ( 955755 )
      Or put dots on your face with a magic marker. Subgraph Isomorphisim is in a class by itself in complexity theory.
    • Re:Hrm. (Score:5, Funny)

      by Tablizer ( 95088 ) on Saturday March 25, 2006 @09:38PM (#14995876) Journal
      Impressive, but what if I shave?

      It is okay, Osama, you were passed up by two presidents, hundreds of Preditor flights, and thousands of US soldiers. I doubt some geek toy will do you in.
    • Even worse... (Score:3, Informative)

      The article implies that you have to be enrolled using this system in order to it to be used to verify your identity. So it isn't any use in finding those that there is a high resolution photograph of without the 2d pattern projected on them to generate the 3d surface. This is only useful for proving that someone is carrying their own valid document, not for picking known criminals out of crowds.
      • So it isn't any use in finding those that there is a high resolution photograph of without the 2d pattern projected on them to generate the 3d surface. This is only useful for proving that someone is carrying their own valid document, not for picking known criminals out of crowds.

        Even if the "known criminals" have been previously arrested and had the procedure carried out (along with "mug shots", finger prints, DNA samples, etc) the article says nothing about the effectivness against people intentionally
  • Great (Score:5, Funny)

    by mcc ( 14761 ) <amcclure@purdue.edu> on Saturday March 25, 2006 @09:17PM (#14995799) Homepage
    I can't wait until the day when I get punched in the face, and suddenly I can't use ATMs anymore.
    • by Anonymous Coward on Saturday March 25, 2006 @09:32PM (#14995852)
      Why wait?
    • I can't wait until the day when I get punched in the face, and suddenly I can't use ATMs anymore.

      Then stop mooning the football team in your sister's chearleader custome.
    • Actually, I *can* wait for the day I get punched in the face.
    • Fcial recognition relies on probability matches, not positive identification. it's just a filter, of varying sensitivity. even being punched in the face or gaining ten pounds might not invalidate your entry. it scans off areas that aren't very suceptible to change with weight or aging. The top of your eye socket, the distance between your pupils, the corners of your mouth, your cheek bone... don't think your going to walk up to an atm and take out cash, you're going to need your implantable rfid and a facia
      • The distance between my pupils you say? Well dang, that's hard to modify. Try this. Focus on a remote mountain, tree, house, whatever. Then you stick your finger up in front of your face. Focus on it. Notice how your eyes move? Yes, the distance has now shrunk. Yay.
  • If this wasn't Rolland, I would RTFA. However, it sounds odd to say an exact 3d image. What is so exact about it? It's a digital representation of reflected light, not a clone. Sounds sort of like saying "an exact photograph" or "a real live stuffed animal."
    • Well, the way I'd look at it would be something akin to how we can accurately get a 3-d image using sonar, and that's only some form of energy projected in one direction. Why not project another, faster form of energy - such as light - instead of sound to get a better, higher-resolution image, at a faster rate?
      • Why not project another, faster form of energy - such as light - instead of sound to get a better, higher-resolution image, at a faster rate?

        Except that a faster form of energy will make actual depth readings less accurate, for obvious reasons. Slower is better in this case, as long as its not so slow to accumulate motion blur. Anyhow, FTFS it sounds like this thing works using computer vision techniques, rather than fancy laser range-finding.
  • by Matts ( 1628 ) on Saturday March 25, 2006 @09:26PM (#14995831) Homepage
    I hate reading things like this - the marketers got hold of it and decided that airport checkins were clearly a problem, and this would help. Sorry, but that's BS.

    Airport checkins take time because they are a security procedure. The "recognition" part of it takes seconds as it is - just swipe the passport or other form of ID. What takes the time is confirming that the traveller's luggage hasn't been modified, finding a decent seat on the plane, and labelling up the luggage they drop. I've never been held up because they couldn't figure out who I was. Ever.
    • Haven't you been keeping up with the program? The only way to get research grants these days is to tie your research somehow into "the war on terror", no matter how tenuous that connection might be.
    • It's not about airport checkins. That's just the palatable "face" (hur hur) of the technology.

      The real use of this is to identify you in locations where you are not presenting your ID ; in the street, in shops, on public transport.

      We already have a reliable biometric system for passports and other photo-ID. It's called a "photograph". Making a machine do the work only makes sense if you want to do a lot more of it.
  • Zit! (Score:4, Funny)

    by Tablizer ( 95088 ) on Saturday March 25, 2006 @09:29PM (#14995839) Journal
    So if you get the Mother of All Zits on travel day, you walk/swim to Hawaii.
  • Exactly what this world needs, a 3-d replica of my face.

    So much for us fuglies living in relative anonymity.
  • Modifiable (Score:3, Interesting)

    by TheStonepedo ( 885845 ) on Saturday March 25, 2006 @09:38PM (#14995879) Homepage Journal
    This is a very modifiable "biometric." I lost a little piece of my nose in a nasty bicycle accident. Some people get facelifts, nosejobs, and botox injections. Many men have differing amounts of facial hair on a day-to-day basis. People who fly infrequently could gain or lose a good deal of weight between flights and have different facial dimensions.
    The error tolerances that would have to be built into an automated face scanner would have to be large. I would rather have a human check my ID in a few milliseconds more than have an inaccurate system for verification. Show me a 40 millisecond thumbprint scanner with an international database and we'll talk.
    • Re:Modifiable (Score:2, Interesting)

      by radiotyler ( 819474 )
      Speaking of that, I'd love to see some folks in the special effects industry play around with these biometrics systems to see how easy it is to spoof someones face, or invalidate their entry in the database with make-up or whatever magic they use these days. Why are there never any MythBusters around when I need them?
    • Why fingerprinting? You can burn off your prints and there are ways to spoof them. Let's go with iris recognition. It worked in Minority Report, and it should be fairly effective for the time being as you can't get an eye transplant for a couple grand yet.
  • Camera (Score:3, Funny)

    by HermanAB ( 661181 ) on Saturday March 25, 2006 @09:39PM (#14995881)
    Sounds like Kodak moment to me... :)
  • at best, incremental (Score:3, Informative)

    by penguin-collective ( 932038 ) on Saturday March 25, 2006 @09:40PM (#14995886)
    3D-based face recognition has been tried before, and you can buy 3D scanners that use projected light patterns commercially. So, there isn't really anything particularly new about this.
  • Now having my passport stolen will be the least of my worries...
  • We can't build a working, reliable transport to earth orbit in 60+ years, but we can build superscience security in less than five years. Guess we really want to lock ourselves into prison. Not enough spirit or imagination to create a way out of the jail we were born into.
  • Goodnight, privacy. (Score:3, Interesting)

    by TheNoxx ( 412624 ) on Saturday March 25, 2006 @09:50PM (#14995911) Homepage Journal
    You know, the similarities between the growth of technology and the "fictional" world in the book "The Traveler" [amazon.com] are getting a little too eerie for me.

    I wonder how long until people start using diluted steroids and such to temporarily alter their face's shape (as mentioned in the book, of course) to get around the rapidly advancing face recognition technology, for good or evil motives?
    • It is a good thing I took my kids to this theatre installation today:

      De Hoofdenfabriek (The Head Factory ) Anneke de Lind van Wijngaarden / Aart Jan van der Linden / Dr. Hackmesch

      Perhaps you've also grown a bit tired of your own head and are ready for a new model? Then simply go quickly to the Head Factory and have your head remodelled! Just figure out who you would rather be, and tell it to the doctor! The world-famous specialist Dr. Hackmesch has come over from Germany especially for the Tweetakt f

  • works in the dark (Score:5, Interesting)

    by twitter ( 104583 ) on Saturday March 25, 2006 @09:52PM (#14995915) Homepage Journal
    I wonder if an infrared pattern would work. You would not even know you had been scanned.

    • We probably generate too much IR interference from body heat. Ointments like IcyHot also make a big difference to IR cameras. Any projected grid strong enough to punch through that interference would be noticably warm.
    • It certainly would, from which one can draw the inevitable conclusion that the only way to escape from Orwellian, Brazillian horror is to annihilate all the humans who don't share your genome with a timed superspreader virus.
  • This technology should speed airport check-ins

    Just one more reason for me to continue to not fly. When are people going to realise that catalogue and tracking every movement of every person is not only an extreme invasion of privacy but a pathologically insane thing to do?

  • by Anonymous Coward
    If faces are made illegal, only criminals will have faces!!!
  • by JimBobJoe ( 2758 ) <swiftheart&gmail,com> on Saturday March 25, 2006 @10:05PM (#14995951)
    but it could also be used in banks

    I believe the biggest problem banks have is ordinary robberies. I can't say I've heard of any situation of someone using someone else's identity to wipe out their bank account.

    Even in countries that do not have identification cards (and, after all, the english speaking world fell into this category until only fairly recently) I haven't heard anything to suggest banks are having/have had troubles identifying customer's correctly. (The identification collected when opening an account now is for Patriot Act purposes.) The lowly ATM with 4 digit pin is used successfully without identification (phishing is its main weakness.) Thanks to debit/credit cards, identification is now even less relevant to banking.

    Why's this crap always being pushed on banks?
  • A pattern of light is projected on your face...
    And you are recognised within 40 miliseconds.

    So I guess this pattern of light would appear like nothing more then a brief flash to a person.

    Really, someone's been watching Minority Report too many times.
  • To watch, identify, and track people as they mill around on the side walk minding their own business, commiting no crime. Just cant allow people to live their daily lives anonymously can we.
    • I know this might not be a popular question, but why should someone be able to walk down the street anonymously? all our tax dollars paid for tht street after all... I draw a line for privacy, and it's called society. I don't think there's much of a right to operate within society anonymously. Noone's going through your pockets or pulling your kids out of bed. They just want to know who the guy going into the grocery store. of course technology like this would be a severe deterrant to crime.. but who cares.
      • You should be able expect a certain level of anonymity out in public, as a law abiding citizen. Sure you can be seen by others, and observed, but that is far different then being identified, tracked and logged.

        Just because your actions are 'innocent' and 'I dont have anything to hide' doesnt give anyone the right to investigate you. Its none of their business.

        Oh, and remember i paid my part of that sidewalk too. And we are not talking about a 'dumb' security camera stuck on a wall here.
  • how does the system respond to these?

    existing face biometrics systems break down with simple alterations like those iirc...
    • how does the system respond to these?

      "Please take your glasses off, sir. Thank you."

      Actually, dealing with beards is probably easier than you can imagine. If the person has a fixed beard style it won't change much between scans, plus, the beard doesn't cover ALL your face. Only the surroundings.
      • heh, ok. but isnt there a risk of false positivs or false negatives given that you get less "accurate" data?

        if one wear a full beard, everything from the cheekbones down is coverd. to me that sounds like very little data to provide a accurate biometric profile from...

        still, this will mostly be used to look for islamist terrorists right? so who cares as they are the only ones wearing full beards these days...

        basicly there is to much fluff and to little substance in all this...
  • I know that things like the patterns in our eyes, fingerprints, and the blood vessels in our faces are never supposed to repeat in somebody else, but what about the shape of a face? If they build in the tolerance for weight gain/loss, facial hair, etc. could somebody else have a face similar enough to get through? What about identical twins?
  • by Symp0sium ( 961148 ) on Saturday March 25, 2006 @10:48PM (#14996097)
    My twin brother is going to empty my bank account now! thanks Sheffield Hallam.
  • by NorbrookC ( 674063 ) on Saturday March 25, 2006 @10:49PM (#14996098) Journal

    FTA: giving a digital mapping of a face that would form part of a fool-proof security system.

    I have yet to hear of a "fool-proof" security system of any sort. I've heard many security schemes touted over the years as being "unbreakable" or "fool-proof", and yet somehow someone manages to break them or fool them. This is what happens when PR hype takes over instead of substantive information.

    This is an interesting technology that may have applications down the line, but it's still new, hasn't been given a wide range of testing, and appears to be something that can be spoofed by facial changes. The PR hack at the university needs to switch to decaf.

  • Plastic surgery in Central America. It's cheap there especially Costa Rica.

  • Biometrics used to mean a robber would want my finger for the ATM. Now they want my head?
  • Security? (Score:2, Funny)

    by bmecoli ( 963615 )
    Screw security, I'm going to hook this device up and make a model of my face so that I can play as myself in *insert modern FPS here*
  • I, for one, welcome our new biometric overlords.
  • by nick_davison ( 217681 ) on Sunday March 26, 2006 @12:27AM (#14996396)
    An accurate 3D model of a human face can be constructed in 40ms?

    Excuse my whilst I almost jump up and down with glee. I mean it's not as if a typical high res photograph can be taken in 1/300th of a second (given decent light) and a bunch of them can't be taken simultaneously with a bank of cameras - leaving almost all of the remainder of 1/25th of a second to quickly calculate a 3D model using the same digital photogrametry that's been around for years on a powerful enough system.

    To put it in context, there have been camera systems that can film an actor "in 3D" - and then use that co-ordinate data to manipulate a 3D character - for TV use for the last half decade or so. By definition, at 25 frames per second, it too builds a 3D model within 1/25 of a second (40ms). The only difference is higher accuracy.

    So, OK, they've come up with a new technique for projecting a dot pattern that makes it even easier to record a set of points than the old annoying stick on black dots method. Even so, quickly capturing a 3D image isn't radically new - a bank of cameras can capture it in far less time than 40ms and you can do the processing in a staggering 2-3 whole seconds as the person steps away before the next person steps up.

    The slow bit has always been comparing a complex 3D model against a huge database and identifying matches when people move their facial muscles between each image.

    Of note is the simple fact: This talks about how "fast" 3D imaging is now available (although it has been by years but we'll ignore that) which is a totally different concept to actually comparing that information against several million, if not hundreds of millions, of other entries.

    OK, so fair enough, the article talks about comparing someone to a specific record to see if they're who they claim to be. Again, nothing that couldn't already be done with a bank of decent CMOS based cameras. They imply that this is "more" accurate (which I still dispute is any more accurate than a bank of 10MP CMOS cameras and traditional photogrametry) but make absolutely no reference to cracking the real problem of people's 3D facial structure changing as their muscles move, as they gain weight, etc.

    So - they have a quick method of creating a 3D snapshot, which could be done already, and haven't solved any of the real problems that make a simple 3D snapshot useless (comparing against large numbers of possibilities, parameterizing 3D points that move as faces do). So absolutely nothing then? Sweet.

    I wonder if I could come up with a new, different, but absolutely no better technology for something people also still can't do very well. If so, I wonder how much a breathy press release and vapid article would net me in grants for my research?
    • A bank of cameras is more expensive than a single, 2D camera and a fast processor. This is an application-type advance, not just an ability-type advance. This means that it'll be easy to do outside of a research situation. Your "solutions" that you present are all expensive, cumbersome and error-prone. This has many fewer drawbacks, so it's more likely to be rolled out.
      But then again, you know better than anoyone else. Of course. You post to /., how could you not?
  • 3D Faces - The Movie (Score:2, Informative)

    by 0x15 ( 852429 )

    If this interests you, MERI has additional information in the form of a movie [shu.ac.uk] about it.
  • by aepervius ( 535155 ) on Sunday March 26, 2006 @03:19AM (#14996811)
    Here is it :
    Combine together two of those :
    * iris recon
    * 3d face recon
    * fingerprint
    with one of those :
    * a pin code
    and one of those :
    * a secure card with a chip and a recent encryption technology

    As it is always said here on /., something you are, something you have, something you know.
  • Hello, first post and all. Saw this news article and it made me laugh. 40ms for taking the picture, maybe, but that doesn't include all the other time involved. I'm a student at Sheffield Hallam and I've been taught by the lecturers involved. What's more I've had my face scanned in. I can tell you that 40ms is very, very deceptive. So maybe it does take 40ms to take the photograph but it isn't a stunningly high resolution photo and even then it is only a photo. The system works by taking a normal photogra
  • I don't know about these guys' system but it isn't a breakthrough. Unless you count total loss of public anonymity a breakthrough. Anyway I've seen videos from a structured light scientist, and this is published at least in ACM Siggraph, of what I believe is 60Hz 3D reconstructions using a video camera. Yes that is 16.6 ms per frame. Or if you want 2 frames you are still below 40 ms. Looking for the guy's name again but google for "structured light video camera 3d reconstruction 60Hz" and you get a bu
  • First of all this seems not very novel to me. They already aquired a patent for this technology int he U.K. in 2004. I am far more impressed by the abilities of the eyeQ system by Mobileye [mobileye.com], a intergrated camera with processor being able to perform complex image processing. It would not surprise me, if the eyeQ could do this job as well. The image processor is RISC based, and although it only runs as 120 Mhrz its computing power is theoretical equivalent of an Intel Pentium IV processor, running at 4Ghz cloc
  • It will take nearly an hour (over 53 minutes) to process the individual faces of people packed into an 80,000 seat stadium to watch something like the superbowl.

    This is assuming you can even isolate each face from all the others (and that the camera doesn't linger on Janet Jackson's torso during the half-time show...)

To do two things at once is to do neither. -- Publilius Syrus