Stories
Slash Boxes
Comments

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

Face Recognition - Real or Science Fiction?

Posted by CmdrTaco on Wed Oct 25, 2006 09:42 AM
from the oops-i-did-it-again dept.
An anonymous reader writes "Facial recognition software has been touted as one of the technologies that will change our future, particularly in law enforcement. How close are we to being recognized by a computer anywhere we go, as portrayed in movies like Minority Report? According to the industry's recent Public Relations releases, these products are closer than we think. The reality though, is that current products work only when utilizing a small comparative sample, and any attempts for an individual to disguise themselves typically throw off the results. To see how far this technology needs to go before becoming mainstream, one site utilized Government-tested face recognition software, available freely through MyHeritage.com, to compare hundreds of famous people, animals, and cartoons to a database of 2,000 celebrities. Some of the results showed promise for the technology, but most were just funny — for example, who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?"
+ -
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This discussion has been archived.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 Lurker2288 (995635) on Wednesday October 25 2006, @09:44AM (#16577946)
    "who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?"

    I think it's more a question of 'how many beers' than of 'who.'
    • Re:trick question (Score:4, Insightful)

      by WindBourne (631190) on Wednesday October 25 2006, @10:36AM (#16578824) Journal
      Or what computer program, camera, and lighting that you have. And add facial hair (perhaps on all), cosmetics(again on all), or even haircuts. Basically, it will always fail on those that do not want to be recognized. But down the road(20 year), it will work well on those that are not suspecting it i.e. it will be a good way to track down regular citizens when the government is granted the power to grab whoever they want and attribute it to say terrorism. Fortunately, we are a long ways from that. Or are we?
  • by 2.7182 (819680) on Wednesday October 25 2006, @09:45AM (#16577960)
    After working in computer vision for 5 years I've realized that most problems aren't hard - they are not well defined. Mathematically face recognition is not a problem that can be stated.

    Many other problems in CV are like this - edge detection, segmentation, etc. But people write hacks that work in restricted conditions and say they've solved.

    And look, you could always just put on those Groucho Marx glasses.
    • by PieSquared (867490) <isosceles2006@gm ... 926.com minus pi> on Wednesday October 25 2006, @10:08AM (#16578346)
      I think the real problem is what it looks at. The shape of your face is what it looks at. What if you put a little clay or really thick makeup around your jaw and cheek bones to change your visible facial structure... and of course facial hair can be shaped to look like pretty much anything is under it without even adding anything artificial to your face. And of course, you'll need multiple frames of reference with a wide angle between them to get any useful information anyway... you can't really judge depth from a single frame and if you try a little eyeshadow will throw it off. I can only see facial recognition as proving that you aren't someone smaller then you are, not that you are a specific person. And of course you could always get one of those masks from mission impossible! Yea, that's what I thought!
      • Here are two [wikipedia.org] pictures [comedy-zone.net] of Matt Stone, one of the pictures of South Park, just to prove your point.
        • Re: (Score:3, Insightful)

          Are skulls that unique? I was sure the unicity was contributed to by not just bone structure, but muscle structure, cartilage(the nose) and pigmentation. In fact, do we know for sure that no two faces are unique? Until facial recognition can tell fraternal twins better than a human can, perhaps we shouldn't put those in mission critical environments, shall we?
    • by King_TJ (85913) on Wednesday October 25 2006, @10:23AM (#16578628) Homepage Journal
      But don't we almost always get a computer to solve a problem that's not strictly a mathematical one using "hacks that only work in restricted conditions"?

      Our spell-checkers in our word processors don't actually know anything about the rules of a language, phonics, etc. They just do lookups from a dictionary. If a word's not listed, it has no idea if it's spelled properly or not -- even if the misspelling is one that's simply not a possible correct sequence of letters for the language. Most don't even realize if a word is misspelled in the context of the sentence, as long as it matches a correct spelling in the word list.

      Until we figure out how the human brain recognizes faces as individuals, we can't expect anything *but* a clever hack for a computer to do the same. And truthfully, I suspect the human brain takes many things into account to do a "recognition" on a person. How often do you see somebody in the store that you're pretty sure you know from a previous job, school, etc. but you're not quite sure? I've had this happen a few times, and to make a better determination, I had to take other factors into account, like the sound of their voice if I heard them speak, the way they walked, or maybe an expression that came across their face. Humans "key in" on specific things that help them remember a person. And depending on which "features" they chose, they may or may not be effective. (Say you remember a gal really well because of her long, flowing hair? If she cuts it real short, there's a good chance you won't recognize her at all anymore if she walks by you.)
    • by Illserve (56215) on Wednesday October 25 2006, @10:48AM (#16579048)
      Many other problems in CV are like this - edge detection, segmentation, etc. But people write hacks that work in restricted conditions and say they've solved.

      Having worked in brain science for years I can say that the brain itself is a collection of hacks.

      It's just a very huge collection that covers all of the bases that we find ourselves in from day to day. Put a brain in a situation it's not designed to handle and it breaks down just as badly as many artificial CV algorithms do.
      • by lawpoop (604919) on Wednesday October 25 2006, @12:41PM (#16581312) Homepage Journal
        Brain hacks seem to be fundamentally different than computer hacks. Or, the brain seems to have a collection of hacks that we have almost no understanding of, in addition to the hacks that we do understand.

        Ever since the advent of solid state electronics, it was said to be only a matter of time before robots would be sweeping, washing dishes, performing surgery, etc.

        Things that we think are really simple, that even retarded people can do, like recognize a face or a voice, understand speech, move bipedally with grace (hell, with any number of legs -- 2, 4 or 6), pour a glass of water, etc. are *hard* for robots and AI. We don't even have a model for how these things work. Even really dumb animals like turkeys can run through their environments and successfully hunt and catch flying insects.

        We do have robots that are getting good with articulation, like Asimo, but we still aren't sure whether they are using the same 'tricks' that organisms use. That is to say, they are a solution to the problem of bipedal motion, but we don't know if they are the same solution that the human mind is. I'm not sure that we have even a model of what solutions organisms use.

        Meanwhile, things that we think are difficult, like playing chess, factoring polynomials, or other kinds of difficult math, are easy for a computer. Now we know that the brain can do complex math like trigonometry, in order to accomplish tasks like catching a ball. but that doesn't help the average person play chess or do complex math on paper. However, the average person excels at these hard AI problems, like having a conversation or pouring a glass of water.
    • Re: (Score:3, Interesting)

      I also worked on a project to compare images. The idea was general purpose, not just for faces only. Unfortunately, it didn't really work. I took the woefully slow prototype and rewrote it so it worked correctly, and was far faster. And all that did was help destroy their illusions and delusions that it was going to work. Before, they could be optimistic because it wasn't fast enough to do hundreds of tests, and so they were able to point to extremely small sets of data upon which it had apparently mo

    • I'm afraid I'm going to call shennanigans on some of this. I've been doing Vision work for about 5 years now with a hefty does of image and signal processing in the mix(Working as gradstudent in the field right now in fact). Edge detection is well defined. The canny and shah-istan(think that's the name) are about as close to a mathematical optimal edge detector as one can get. There is in fact a well developed body of theory regarding differentiation of Signals. The problem doesn't lie in the mathematic
  • recognized (Score:4, Interesting)

    by AcidLacedPenguiN (835552) on Wednesday October 25 2006, @09:46AM (#16577982)
    This is all well and good, but the minute I get falsely identfied as a criminal just for being in the bar district late at night in the wrong place/wrong time I won't be too happy. . .
  • so I guess... (Score:3, Insightful)

    by theStorminMormon (883615) <theStorminMormon.gmail@com> on Wednesday October 25 2006, @09:46AM (#16577988) Homepage Journal
    So I guess next time a teletubby or Shrek wanders through a mall, they're totally going to throw off the face-recognition software.

    Is it just me, or does that seem like a stupid way to test the software? If you want to show that rudimentary disguise is an easy way to get around it, that's valid, but just messing with the sample of potential matches by throwing in cartoon characters destroys the validity of the "study".

    -stormin
    • . . .Though I'll still laugh when Streisand gets framed for robbing a bank or something. . .
      • . .Though I'll still laugh when Streisand gets framed for robbing a bank or something. . .


        But how are you going to convince Shrek to rob a bank?

        -matthew
    • Report to PHB version 1:- This software is so poor it threw up many false positives
      PHB :- This software get's positive results. We'll buy it!

      Report to PHB version 2:- This software is so bad it confuses Barbara Striesland and Shrek
      PHB :- My kid wouldn't get those two confused. We won't buy it.

      A 'good' report depends on it's audience. For most /.ers we would want to inspect the data and see a proper statistical breakdown of the results. To catch the public's attention, however, you need to add a few cele
    • "Shrek" might be in a store's window display, as a cardboard cut-out. So when the Feds are looking for Barbara, they may get "false positives" from mall video surveillance all over the country.

      Just saying...
      • ...you might rethink your position.

        Especially when they start the interrogation with: "So we've finally caught up with you at last, Yosemite Sam [wikipedia.org]."

        If the system is designed to match faces against a database of photographs of real humans and you plug cartoons in there - what do you expect? False positives generated in this manner prove nothing about how the system would actually work. If there are a large number of false positives even when the match database is realistic, then we have a problem.

        -stormin

  • But I thought (Score:3, Informative)

    by xirtap (955611) on Wednesday October 25 2006, @09:47AM (#16578000)
    I thought they used chips in the eyes of people in minority report, not face recognition.
    • Re: (Score:3, Informative)

      I thought they used chips in the eyes of people in minority report, not face recognition.
      Retina recognition, I think.
  • MyHeritage site (Score:5, Informative)

    by LoverOfJoy (820058) on Wednesday October 25 2006, @09:48AM (#16578014) Homepage
    I've tried out the software and it was fun for some laughs. I'm not sure how it works exactly but I can tell that the angle of the face makes a difference. When I put one picture of myself in where I'm looking ever so slightly to the right, I'm matched with celebrities photos looking in that direction. When I put in a similar photo facing the other direction, I get a different set of celebrities looking in the other direction. There's a few overlaps and those are the ones I think I look the most like (although it's a stretch to say I have anything that could pass as a celebrity look).
    • Evidently, I look a lot like Tom Cruise. Well, actually, the first hit was Matt Stone [wikipedia.org], but with a much less flattering picture. Tom Cruise was 4th or 5th on the list, after Luke Wilson. Here was the picture [virginia.edu] I submitted. Naturally, I'm being sarcastic when I say it's flawless.
    • Re: (Score:3, Insightful)

      I saw the same thing. Also, if the person in the image is doing something with their face (smiling, open mouth, wide eyes) it tends to match with images of people doing the same thing. Kinda simplistic, more like a trick than a tool.
      • Oh, I have a great idea. If they ever start making wide use of this technology in public places (I'm looking at you, England), all you need to do is smile all the time. Chances are that if you are matched with someone, it will be someone happy, and therefore, probably not a criminal (how many criminals smile for their mugshot?)

        -matthew
  • Inevitable. (Score:5, Insightful)

    by Lethyos (408045) on Wednesday October 25 2006, @09:49AM (#16578034) Journal

    Not to nitpick excessively, but you could easily substitute portions of this article with terms like (and relating to) “Internet”, “personal computer”, “telephone”, “car”, and others. Asking ourselves if a technology is “real or science fiction” when it already exists (albiet in a primitive form) is silly. Of course it exists; the question itself cites examples. Perhaps the meaningful questions might be along the lines of: “what are the challenges associated with making it accurate?” or “what impact will facial recognition have on society?”

  • I'm wondering about the legality of all this, especially in a criminal justice system. My DNA, for example, can't be used in court as evidence unless certain hoops have been jumped through; the prosecutor needs a reason to obtain a DNA sample and then procedures must be followed.

    I wonder if the same systems will apply to a computer analysed image of my face; will there be a criterea for when this image is admissable in court? Will I have rights concerning my image? Or are we just going towards a 1984 styl
    • Re:Legal hoops (Score:4, Insightful)

      by Quadraginta (902985) on Wednesday October 25 2006, @10:17AM (#16578512)
      Using a computer-captured image of your face in Court would presumably come under the same rules as using a photograph of your face. More or less, if you appear in public, your image can be used.

      The more interesting question, I suggest, is whether a computer recognition of your face is going to be in any way equivalent to a human recognition of your face.

      For example: if you stroll into a 7-Eleven, and the donuphage with a badge sitting there swilling coffee thinks you look like a famous bank robber whose mug has been circulated by the FBI, then he's entitled to take you into custody, and search you (for his own safety and those nearby, et cetera). If he finds half a gram of coke on you, you're in trouble. Now suppose it isn't the cop's eye/brain combination that "recognizes" you as a bank robber, but rather his shoulder-mounted camera/computer combination. Is he still entitled to act in the same way?

      You can argue it both ways: (1) the camera/computer is almost certainly always going to be worse at this kind of thing than the eye/brain. Recognition is about the single most important thing our eyes and brains do, and they are highly optimized for it by natural selection. If it could be done better and faster, we would do it. So, we should trust the camera/computer less. But (2) the camera/computer is not subject to the vagaries of human psychology, mood, et cetera. The cop may take you in unreasonably because he doesn't like your skin color or length of hair, the camera/computer isn't subject to the same prejudices. So maybe it's better to trust the mindless device.
      • Re:Legal hoops (Score:5, Insightful)

        by B3ryllium (571199) on Wednesday October 25 2006, @10:30AM (#16578728) Homepage
        or, maybe it's better to not carry a half-gram of coke on you.
      • I would assume that this software is really just a way for computers to sift through lots of photo/video to look for possible matches - it is then the job of humans to make the final call. Even humans make false positive id's, so we can hardly expect computers to ever be better at judging such a subjective and changing real world feature as the human face.
    • They're not trying to use this to prove you are one person or another in court.

      If you have a surveillance video of a crime scene, and a huge database of faces of citizens, then you could use this to narrow down your list of suspects. Obviously once it ever gets to the court system, there will be plenty of human experts on hand to look at the camera footage and testify as to whether it looks like you.

      Also, this could be used to flag "suspicious" people at airports and other places. That way, the human securi
    • IF by "certain hoops" you mean "anyone ever arrested, regardless if they were ever charged", then yes.
      http://www.washingtonpost.com/wp-dyn/content/artic le/2005/09/23/AR2005092301665.html [washingtonpost.com]
    • Why? They have to get the warrant to get your DNA. Once they have it, they can test it.

      Getting a picture is easy. In fact, they take one when they book you. Why shouldn't they be able to use it?

      Also, they use non-scientific face matching all the time (line-ups and eye-witnesses). Why shouldn't they be allowed to do the same thing with a more accurate and unbiased judge like a computer over a person (who may be the victim, or biased in some other way)?

      I would think it would be allowed now without any kind

  • by Lisandro (799651) on Wednesday October 25 2006, @09:52AM (#16578092)
    For example, who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?

    So, i see it's working correctly!
  • Humans seem to have no problem with it, so it's clearly possible.
    • It's perfectly possible to get very good results on comparatively small databases (a few 100 faces). See the comprative tests conducted on the Surrey XM2VTS database, or the FERET test runs. These tests don't adress the disguise issues, but do address natural variation over time (breads, hair-style, presence or absence of glasses). Partly because computer vision is still a comparatively young field, there's very little work on performance of long time periods. Most databases cover only two or three years

      I

  • Isn't this the sort of problem that might be well-addressed with genetic algorithms? We've got a problem that we can't really define well (mathematically), but we know what good results should look like. So the actualy process to solve the problem is hard to design, but a test for good results is (comparatively) easy to design.

    That sounds exactly like the sort of problem that you could use GAs on.

    Unless, of course, I'm completely wrong about the state of the art in genetic algorithms, or am making some othe
    • Sounds good. You could use GA guided neural networks. Worked once (i.e., us), so it stands to reason it could work again. I'd set up an input system similar to how we understand V1 to work, and then let the GA explore over a range of options for V2, V3, and V4. Each of these would be constrained by some of our best guesses of how they work in humans. I've been trying similar things recently with a model of the hippocampus [neurojet.com].
      • Re: (Score:3, Interesting)

        The problem is the inputs. Do you inputs sets of geometry (eyes are X" apart, at an angle of 0.53 degrees, chin is .5" below lips, blah blah blah), the raw image, or something else? If you use the raw image, you'd need a system in the front end scale/rotate the images to be in about the same place otherwise you probably have no chance (unless you want your neural net to do that TOO, which would make training harder and take longer).

        Even if you use geometry (we have a vague understanding of what makes peopl

        • If I was training to match V1-4, I'd have the input come from two "eyes" with inputs similar to what our eyes actually provide to our brain. We know quite a bit about visual cortex, but there's a lot we don't know. Initially, I'd train it using a batch of photographs for a single person (we'll call her "Momma") and then I'd train with a few others (where a match is a match only if it's the same person). From there, I'd create histograms of parameter settings that seem to do an adequate job on this small set

    • but a test for good results is (comparatively) easy to design.

      Actually, it's terribly hard. You can say that "this picture should match this person and that picture should match that person", and that will give you an algorithm that works great on those faces, but can/will give horrible results on everything else.

      Of course, if your genotype representation is good, a GA would surely do wonders, but then you're basically back to where you started, how to represent and interpret a face in general.
  • Cool (Score:2, Insightful)

    Okay, so I look like your regular geek. I'm fat and bald. Yet I apparently look like some hot female movie stars. 80% like Grace Kelly. Not sure if I can link directly to a result. Let's try: http://www.myheritage.com/FP/photo.php?siteID=1&p h otoID=5969307&source=album&sourceID=963790&albumID =963790 [myheritage.com]

    Granted - the examples looked pretty good, but I just can't see Grace Kelly when I look myself in the mirror.
  • Here to Stay (Score:5, Insightful)

    by DumbSwede (521261) <slashdotbin@hotmail.com> on Wednesday October 25 2006, @10:11AM (#16578376) Homepage Journal
    I believe Minority Report used retina scans, but that nit aside facial recognition works to a degree and will only get better. Security cams will eventually upgrade to HDTV resolutions, perhaps augmented with very high resolution stills when a potential match is made. This will all take more processing power, but all mighty god Moore will eventually gives us this day our daily CPU load.

    About false positives. So what? Eyewitnesses make mistakes also. Eventually, perhaps very soon, machines will surpass humans in this arena just as they have in others. Can anyone here on Slashdot defeat Deep Blue at Chess?

    As to the legality or ethics, what can be done will be done, at least in public areas. If it would be legal for a human to do (they haven't outlawed humans scanning for suspects in public areas) then it will be legal for machines to do despite the unease many will feel knowing they are constantly being watched.
  • by MobyDisk (75490) on Wednesday October 25 2006, @10:15AM (#16578466) Homepage
    Some of the results showed promise for the technology, but most were just funny -- for example, who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?"
    That's just trolling. The software was instructed to find the celebrity who most closely matched a cartoon character. It didn't mistake anyone for a cartoon character. And since cartoon characters are not within the scope of what the software is for, it shows that it worked better than expected. Attempts like this to belittle the success of the technology are akin to Ad Hominem attacks, and have no merit in a discussion.
  • Cartman: Try this on for size. Blood-drenched, frozen tampon popsicle!
    Sadaam Hussein: Hey buddy, I know I was mean before, but don't worry, I can change!
    Cartman: Okay.
    Not. Fuck, shit, cock, ass, dildo, boner, bitch, pussy, butthole, Barbara Streisand!
  • What about plastic surgery? Identical twins? Even a close sibling could be similar enough to fool software in some cases.

    On the other hand, fingerprints are completely unique, even between identical twins, and (last I heard) unchangeable. Researchers would be better off spending time on improving fingerprint-scanning technology for identification purposes, although clearly face scanning would be even less intrusive for other tasks.
  • by creimer (824291) on Wednesday October 25 2006, @10:33AM (#16578776) Homepage
    Every morning I wake up to look into the mirror and it's a different face that I don't recognized. Maybe I need to upgrade my mirror?
  • Showing the resemblance in absolute numbers, better with some P-value or E-value.
  • We simply need strong pattern matching algorithms for images...then anything could be recognizable. The brain uses pattern matching, doesn't it?
  • "How close are we to being recognized by a computer anywhere we go, as portrayed in movies like Minority Report?"
    Now I could be wrong but I am pretty sure Minority report was portraying retinal scanning not facial recognition
    • So you think they should compare Mike's voice with Barbara's face?

      Shrek's face is not Mike Meyers' face. Since they are comparing faces, this is not an invalid comparison, as you think.