Forgot your password?
typodupeerror
Graphics Software Science

Detecting Faked Photographs Gets Easier 258

Posted by michael
from the cowboyneal-with-hat-is-REAL dept.
nusratt writes "Some years ago, an issue of 'Whole Earth' had a convincing cover-photo of a flying saucer cruising low over downtown San Francisco in broad daylight. The accompanying feature article proclaimed that photographs can no longer be trusted as evidence of anything, because of the ease of doctoring images digitally and undetectably. Now, Dartmouth Professor Hany Farid and graduate student Alin Popescu 'have developed a mathematical technique to tell the difference between a "real" image and one that's been fiddled with.' Farid says, 'as more authentication tools are developed it will become increasingly more difficult to create convincing digital forgeries'." There's also an NYT story.
This discussion has been archived. No new comments can be posted.

Detecting Faked Photographs Gets Easier

Comments Filter:
  • Seamless Math Next? (Score:5, Interesting)

    by mfh (56) on Saturday July 24, 2004 @01:11PM (#9789411) Homepage Journal
    While it may become increasingly difficult to forge digital images, and even forge hard currency, the result could be of two possibilities;

    1. Forgers get smart and use older cameras to take a picture of a digital forgery to pass as an original, using blurring techniques offered by physical means and lens... etc (easy)
    2. Forgers quit being forgers (unlikely)
    3. Alteration technologists create armor against image forgery detection algorithms (possible)

    For me, I think any time spent trying to beat the detection of forgeries would be a good thing in terms of art and creativity -- not to mention the possibility of better digital growth algorithms to join layers mathematically seamlessly (which could be used in games and simulation engines for better realism). However, law enforcement agencies might try to combat the circumvention of forgery detection by charging people with crimes for only trying to make their images more realistic and improve technology. It's a messy issue, that will sort itself out over time.

    In Doom 3 Bloopers, a mod I've started on, I am looking at ways of integrating realworld imagery into the mod, and this detection stuff could actually help me to better integrate my own art and images if I can find a way around it. Let's face it, if the math says it's an original, the human eye will be fooled, which is the goal of most video game design. If anyone wants to help along those lines, they should contact me [zenbuzz.org]!
    • by Anonymous Coward on Saturday July 24, 2004 @01:36PM (#9789563)
      I gotta say, while the first half of your comment was certainly insightful, that was one of the most blatant attempts to force your pet project into an unrelated comment I've ever seen!

      You, sir, have balls!
    • by Ieshan (409693) <ieshan@g3.14mail.com minus pi> on Saturday July 24, 2004 @01:45PM (#9789619) Homepage Journal
      It seems as though the algorithm they have works on the fact that there's some statistical difference between a real image and a faked one, but not 100% of the time.

      Why wouldn't someone simply build a filter into a program that changed bits in an image until it passed a check by their algorithm - on failing, it would simply go back and change more appropriate bits?

      Seems as though it would be a computationally intensive but a logically easy task.
      • by yintercept (517362) on Saturday July 24, 2004 @02:01PM (#9789706) Homepage Journal

        My first thought on the article was the same. The mathematical tests to determine fake photos will have value only until the fake photo industry builds the tests into their software. For that matter, I suspect that the main use of the Hany Falid method will be to make your fake photos even more realistic.

        The most interesting line in the article was:

        but computers make it easier for more and more people to manipulate images.

        One could read into these lines that the ability to fake photographs was great until anyone could do it. Now that we know how easy it is to fake photographs, we no longer implicitly trust messages...but we will trust mathematically authenticated fake photographs because math is infallable.

        • by B'Trey (111263) on Saturday July 24, 2004 @02:55PM (#9790004)
          One could read into these lines that the ability to fake photographs was great until anyone could do it. Now that we know how easy it is to fake photographs, we no longer implicitly trust messages...but we will trust mathematically authenticated fake photographs because math is infallable.

          Or you could read into them that when it was rare and difficult to fake photographys, most of the photographs you saw were genuine, so you could place a decent amount of trust in what you were seeing. Now that faking photos is easy and commonplace, you can no longer place much trust in photos. With mathematically verified photos, you can place more (though not complete) trust back in the photo. It isn't foolproof but the level of assurance is significantly higher.
          • by Megahurts (215296) on Saturday July 24, 2004 @03:52PM (#9790307)
            One could read into these lines that the ability to fake photographs was great until anyone could do it. Now that we know how easy it is to fake photographs, we no longer implicitly trust messages...but we will trust mathematically authenticated fake photographs because math is infallable.

            Or you could read into them that when it was rare and difficult to fake photographys, most of the photographs you saw were genuine, so you could place a decent amount of trust in what you were seeing. Now that faking photos is easy and commonplace, you can no longer place much trust in photos. With mathematically verified photos, you can place more (though not complete) trust back in the photo. It isn't foolproof but the level of assurance is significantly higher.



            Or you could realize there's inherent error in any statistical method, and that with a little bitof foresight, this error could be expoited to provide false results. The best detector of falsified photography is still a well-trained human eye. The key to knowing whether or not to trust images is to train your eye.
          • by yintercept (517362) on Saturday July 24, 2004 @06:16PM (#9790991) Homepage Journal
            when it was rare and difficult to fake photographys
            It was never rare nor difficult to "fake photographs." From the very moment photography existed, people realized that they could control the message the photograph carried. Most of the early photographers had studied optics and developed their own film and manipulated images accordingly.

            Most of the early photos you come across were staged. Taking a photograph was a big deal. People dressed up in their best outfits and the photographer would construct the scene. Often the photographer colored or otherwise enhanced the image. Photography has always been an art form. As an art form, people choose the message to convey.

            Using photographs as evidence has always been problematic. The struggle is for lawyers to keep enough faith in photos to be able to use them in court. Personally, I think having a method to declare a photography mathematically correct immediately creates a problem where people with sufficient resources to fake mathematically correct fake photos will have the ability to manipulate the courts.

            There never was that much faith in photographic evidence. I think we are better of having doubts about photographic evidence than we will be if we sanctify any photos as mathematically correct. The methods will have some value in quickly identifying tampered evidence, but will not have value in verifying it.
            • by B'Trey (111263) on Saturday July 24, 2004 @06:46PM (#9791122)
              OK, if we're going to have this conversation, we need to define what we mean by "fake photograph." Yes, many photos were staged, particularly when taking photos was significantly more difficult than "point and shoot." But a staged photograph is not a fake photograph - it's still a reasonably accurate representation of reality as it existed at the time the photo was taken.

              You're also quite correct that photos are routinely manipulated in the dark room. However, manipulating the color and otherwise enhancing the image is not at all what most people mean by a "fake photograph." There's a fundamental difference in those types of manipulations and putting Sarah Michelle Gellar's head on a porn star's body or putting John Kerry and Jane Fonda on the same podium [snopes.com] together. This type of thing was possible before, but it was much, much more difficult and much less common.
              • by angst_ridden_hipster (23104) on Saturday July 24, 2004 @07:25PM (#9791297) Homepage Journal
                At the Getty Museum, they recently had an exhibit of influential photographers and the history of photography (http://www.getty.edu/art/exhibitions/genius/).

                One example photo from the beginning of the last century or before, depicted two figures by a lamppost on a foggy Parisian street. I took it for an unaltered print, but it turned out to be a composite of seven [!] separate images. It was laboriously done in the darkroom. It was incredible. (Unfortunately, I can't find the specific photo on their otherwise excellent site.)

                Now, we've all seen great darkroom manipulation like the work of Jerry Uelsman (www.uelsman.net), but this particular picture was a hundred years older than his work. It was absolutely convincing.

                I guess my point is, photographs have in fact been "faked" as long as there have been photographs.
      • by NanoGator (522640) on Saturday July 24, 2004 @05:15PM (#9790681) Homepage Journal
        "Seems as though it would be a computationally intensive but a logically easy task."

        Probably. Consider this, though: Fingerprints are common knowledge. We ALL know that if we commit a crime, fingerprints will be lifted to try to catch us. There are well known ways to defeat this, but remarkably, LOTS of people are still leaving fingerprints as clues at a crime.

        I hope you can forgive me for reading a little more into your post than you actually said. I don't know for sure if you were going the "this could easily be defeated, thus it is ineffective" route. But I thought this would be as good of time as any to mention this.
    • I think point 3 will come true first, as the truly "skilled" forgers will come to notice a lot of those "real world" factors as light correction (joining two pictures with differing angles to light sources). Layer-fudging is certainly interesting, when it comes to overlap, but I'd be surprised if light and reflection don't also play a role in the algorithm.

      Perhaps some of the forgers(maybe just the ones who used to forge paintings *tongue in cheek humor*) already do pay such attention, after all, attentio
    • by freshmkr (132808) on Saturday July 24, 2004 @02:25PM (#9789852) Homepage
      Let's face it, if the math says it's an original, the human eye will be fooled, which is the goal of most video game design.

      Don't be so sure about that!

      I don't know what these researchers are doing, but I can venture a guess. You can think of an image as being composed of a large set of superimposed sine waves---they look like this [ouhk.edu.hk]. To figure out what sine waves make up an image, you do a Fourier Transform, which is well documented on Google.

      Natural images contain a characteristic power spectrum: some frequencies--lower ones--tend to occur quite often, while others--higher frequencies--are less common. The spectrum is actually pretty regular across an image set. I'm betting, though, that fake images don't respect this power spectrum and lead to detectable anomalies.

      But beware! You can have completely bogus images that also respect the power spectrum. Some researchers at MIT (Torralba et al.) use power spectra to successfully detect different image environments, e.g. indoors, outside in a city, out in the country, etc., but in the papers they show some images that have been reconstructed from their spectral models.

      You would not be fooled. They look like finger paint pictures.

      --Tom
  • Actually... (Score:5, Interesting)

    by julesh (229690) on Saturday July 24, 2004 @01:12PM (#9789418)
    I think authentication tools make it easier. As someone who's tried a little photo manipulation in the past, I can tell you that the hardest thing is knowing when something's right. If you have an automated tool that can tell you when it's right, it becomes easier. Of course, that relies on the tools working...
    • by Anonymous Coward on Saturday July 24, 2004 @01:27PM (#9789516)
      The article says the research is founded by the Department of Homeland Security. That means that despite the many useful possible aspects of this technology, it will probably never see the light of day. If the DHS is going to be using this technology to identify faked photos, it would be greatly in their interest for the full algorithm and its implementations to not be made available to the public-- since after all if people know what the DHS is using to determine faked photos, they can target the algorithm.

      Now that I have written that, looking around, it appears I am actually wrong. If you look at Mr. Farid's personal page, it appears he will be publically presenting [dartmouth.edu] a paper covering the fakeness-detection algorithm. I hope the full algorithm will be presented to the academic community.
      • ... they want to be able to pass off their "intelligence" photos as the real mc coy easier, so as to not get busted when they release crap, like the phony fat osama bin laden video(wicked fake) and the (possibly) berg beheading. this technology would help them to establish bonafides with the offical forgeries. We are *this* close to a running man scenario here with them being able to frame people or to alter public opinion with phony video and pictures. If they can create one and have it slip through the ch
  • Self Defeating (Score:5, Insightful)

    by Dozix007 (690662) on Saturday July 24, 2004 @01:13PM (#9789432)
    While an equation to make sure that a photo is not forged is all well and good, it is self deafeating. Just like simple encryption, which is good for people with simple problems, exploring the equation will yield a way to fool it. If you have someone who truely wants to forge a photo, they probably could still defeat the check.
    • True. And for the same reason, encryption is useless - a determined enough attacker could still break it (assuming he invests enough time / energy / computing power), right? ^_~
    • I don't know what the big deal is. Old style film photos can be fiddled with, but digital photographs cannot be faked, [uncoveror.com] and digital photography is taking over.
    • Re:Self Defeating (Score:4, Interesting)

      by ruszka (456169) on Saturday July 24, 2004 @02:47PM (#9789968)
      Which makes me wonder.. did they create this in order to detect forgeries? or possibly to make it easier for the government to forge photographs and then use the technology to make those images "authentic"
  • by The I Shing (700142) * on Saturday July 24, 2004 @01:15PM (#9789446) Journal
    Now they'll be able to prove that my photo of Rumsfeld shaking hands with Saddam in 1983 is a fake.

    Rumsfeld was really shaking hands with an alien, and Saddam was shaking hands with Elvis, but the resulting merger of the two photos was much more provocative.
  • Just a thought... (Score:5, Insightful)

    by julesh (229690) on Saturday July 24, 2004 @01:17PM (#9789455)
    Farid and his students have built a statistical model that captures the mathematical regularities inherent in natural images.

    I wonder if they've considered the potential applications in image compression?
  • by Anonymous Coward on Saturday July 24, 2004 @01:19PM (#9789463)
    ...but it was only accurate when the photo contained an image of Admiral Ackbar.
  • by multiplexo (27356) * on Saturday July 24, 2004 @01:21PM (#9789477) Journal
    If I take a high resolution TIFF image and start mucking with it I can see how it would be easy to find the manipulations (especially if you're as thumb-fingered as I am with PhotoShop). However with most digital cameras using various compression schemes to store the images how can you tell what is a result of manipulation versus what is an artifact of compression and or digitization? Certainly some gross manipulations will be obvious compared to the properties of compressed images, but I would imagine at some point you'd be hard pressed to say "this image was deliberately manipulated" instead of "this is a compression or digitization artifact".

    While this might not be a problem for gross manipulations (the faked John Kerry/Jane Fonda photo [sfgate.com] being a recent example) I can imagine a class of images where subtle manipulations caused great effects and were not readily distinguishable from compression artifacts.

    • ..and how will it ever work on to prove fake a simple ufo forgery where you just slap something over white/blue background? sure it would probably detect some mild photoshopping made to spice up a picture to look good(or more 'real life', with unnatural blurs & etc). but the more plain the picture is the harder it would be to detect any tempering by just looking at the photo.

      good way to get some pr for some uni though, people like magic like in some bad scifi where they can detect which areas are fake
    • The easy answer to that lies in the original compression artifacts remaining - any new fragment/change will not keep these in a statistically similar fashion, and thats what my understanding of this software is.

      Smudging a part of an image would remove these artifacts, and would be near impossible to reproduce - like the paper grain on a canvass oil painting.
    • by blonde rser (253047) on Saturday July 24, 2004 @02:04PM (#9789721) Homepage
      The NY Times article deals with this directly.

      But Professor Farid said that for now the technique does not work as well with files created in JPEG, the compressed picture format most commonly used online. As the size of a JPEG file shrinks, the correlations between pixels become much less obvious. "At 90 percent quality, it falls apart very quickly," Professor Farid noted.

      So good point. That does seem to be a problem. The NY Times article has more details than the other; it is worth reading.
      • This distinction is important, as it seems unlikely for any means to emerge that will overcome these pixel-level distortions.

        One wonders whether this will lead to a legal distinction between lossily compressed images and others. While audiophiles have long been ape for lossless compression, not as much a need has been felt for graphics. Where do lossless graphic compression efforts stand? Is this an area where a proprietary standard might lead to big $$$?

      • Good point. I should have RTFAMA (Read the Fucking Article More Attentively). Of course someone is probably already working on a Photoshop plugin for "correct faked image distortions", which would be handy to have along with some other plugins such as "remove ex-significant other from picture", "reduce glassy-eyed drunken stare", and my all time number one request, "remove drunk friends standing around with dicks hanging out by your mouth when you're unconscious".

    • I take high resolution TIFF images and "muck with them" every day. Every ad you have ever seen has been digitally manipulated, every video image you have seen. If I am supplied with two good source images, I can create an image that would pass just about any inspection. There are many tricks to get the "irregularities" into (and most times out of) an image. If I wanted to create a fake image, I would do multiple techniques, then "print" the image out to a transparency (slide), or a negative, and have a prin
    • how can you tell what is a result of manipulation versus what is an artifact of compression and or digitization?

      Just a wild idea off the top of my head... while JPEG image compression is a lossy process, it's still a mathematically reproducible process. It might therefore be possible to exhaustively generate the set of originals that could be compressed to the image in question*. If you then apply the detection algorithm to the originals, none of them may pass and you'll still have a valid conclusion.

      *

  • by MedHead (795006)
    I fail to see how an image can have a "random set of pixels" as the article suggests. With color and brightness balancing, the pasted areas of an image blend in with the rest of the image (especially when the artist uses Photoshop's clone brush to combine the two images). This sounds to me like the computer could very well throw out false positives for images that have extreme color or brightness differences.
    • Different digital cameras capture RGB information differently, so, even though on your screen you could be looking at something that appears visually similar, the balance of the Red Green and Blue Channels would actually be different, especially noticeable in the dark areas.

      Another is film grain, not as prevalent in digital cameras, however, with film cameras, grain is varied based on film speed, manufacturer, light level, and other situation. When you comp together images from a variety of sources, one of

  • by Anonymous Coward on Saturday July 24, 2004 @01:23PM (#9789487)
    Farid's algorithm looks for the evidence inevitably left behind after image tinkering. Statistical clues lurk in all digital images, and the ones that have been tampered with contain altered statistics.

    That makes sense. However, it seems like these statistics would be based on very minor details. What happens if you run their analyzer on an image that has been altered by using lossy image compression, such as JPEG compression? [conjoineddreams.net] Lossy image compression is designed to obliterate details humans wouldn't notice; some of these details might be significant to their statistical model. Would JPEG-compressing an image make it impossible to determine its veracity? Would their software just tag all JPEG images with compression below a certain threshold as being "unnatural" (since they have been, after all, digitally altered-- just not digitally altered in a content-relevant way...)

    And don't some digital cameras use lossy formats such as JPEG as their native storage format?
    • Conversely though, I wonder if compression artifacts could be one of the very things the algorithm looks at. Your digital camera takes a picture, compresses it with JPEG and produces subtle (to the eye) artifacts. Now start moving things around in the image (lets assume for the moment you're not even grafting in UFOs), and now the artifacts have moved with them, and can be identified as not being where the JPEG algorithm would put them. Of course, to make it look convincing, you've probably used some smooth
      • Compress your JPEG with three different encoders at subsequently lower quality settings until you reach the quality setting (and file size) you're looking for, and the artifacting will be effectively untracable. Unless there has only been one predictable compression pass since the manipulation, the lossy nature of JPEG should be sufficient to make it impossible to track the chain of encoders back to the source image data. (It's always impossible to achieve the original data, that's what lossy means, but you
        • It would be easier to just use all raw images with no compression artifacts. After the fake photo is complete then it can be compressed if desired.
        • Compress your JPEG with three different encoders at subsequently lower quality settings until you reach the quality setting (and file size) you're looking for, and the artifacting will be effectively untracable.

          In a modern court, the defense doesn't have to prove that you've doctored the picture. If they can prove that you've run the image through three different encoders (rather than coming straight from the digital camera), they will likely establish reasonable doubt that the image is genuine.

  • I've seen picture postcards from the early part of the twentieth century that were obvious forgeries. Things like a picture of grandpa with his six-foot long trout. The Soviets used to erase non-persons (arrested or dead) from published photographs.
  • I wish there was a more detailed description of the algorithm used and/or an online demo available...
  • by bhsx (458600) on Saturday July 24, 2004 @01:29PM (#9789525)
    People making fakes of UFOs and breasts on the Olsen twins will discover new techniques to overcome the detection. It will always continue in a spammer/spam-filter fashion. I'm sure that we'll soon see tools to automatically take the mathematical principles into account and automatically correct fakes that can be detected.
    But hey, I'm pretty sure this one pic of the Olsen twins I got from Kazaa is real anyway.
    • I'd have to agree. It stands to reason that if you can create a mathematical analysis tool that ascertains an images likelihood of fakeness, It is entirely conceivable to use the same mathematical techniques to process an image in such a way that it passes the test.
    • Yes. Just what I thought too.

      If we increase our understanding of image manipulating it will work both ways. If you know how to do something 'backwards' you'll also be in a better position to do it the other way around.
  • by over_exposed (623791) on Saturday July 24, 2004 @01:31PM (#9789534) Homepage
    So if it detects digitally altered images, why can't we just 1) Alter the image to our heart's content 2) Print it on a high quality printer ($500- $2K at CDW) 3) Scan it back in with a nice scanner
    • Print it on a high quality printer

      The short answer is dot structure. All printers, (excepting Dye Sub) use some form of sparying of ink, or layering of screened images(halftone dots) Fancier does not necessarily mean smaller dots, it usually means more calibrateable and more consistent color. What you need is a film recorder, which will transfer your image to a negative or a slide, and currently, there are no digital cameras that will record an image at a high enough resolution for this to be flawless (ab

  • by asterism (148910) on Saturday July 24, 2004 @01:35PM (#9789556) Homepage
    If you are looking for more detailed information, along with equasions, here is a link to one of their recent publications on the topic: http://www.cs.dartmouth.edu/~farid/publications/sp 04.html [dartmouth.edu]
  • I'd make a totally black saucer or triangle shaped balloon, as big as possible on a budget, and fit it with blinking IR and UV spectrum-only LEDs, and then have a few accomplices take photos with both film and digital cameras, along with some other people filming with camcorders on normal, and others with camcorders on Nightshot or equivalent night-vision.
  • by nusratt (751548) on Saturday July 24, 2004 @01:37PM (#9789570) Journal
    NYTimes story link, which is actually more informative and interesting than Dartmouth's own story. In particular, the instant that I started to read the NYT story, I dope-slapped myself for not having thought of the reverse implication of the technology, namely that it might be used to prove that a contraband image (such as child-porn) is NOT faked (and therefore is genuinely illicit).
    • by julesh (229690) on Saturday July 24, 2004 @02:10PM (#9789758)
      No, it couldn't. All it can prove is that if an image is fake it has been done very well.

      If A doesn't conform to the statistical distribution of B, then A isn't B (with a high degree of confidence). But if does, that doesn't mean it is B -- you might just be looking at the wrong set of identifying features. I.e. not everything with two feet and a bill is an aquatic bird; it might be the waiter.
      • "No, it couldn't. All it can prove is that if an image is fake it has been done very well."

        I apologize for having phrased it poorly. My point is the following scenario . . .
        -- Mr.X is arrested for having child porn.
        -- His defense (recently legitimized by the U.S. Supreme Court) is to say, "The images are digital artifice, no actual children were involved."
        -- The prosecutor says, "We ran the images through Farid's process, and there's no evidence of alteration. The artifice which you claim is involved, wo
    • Damn, now somebody is going to have to write a program so that they can process all of their pictures so that their pictures can be plausibly deniable. Maybe not, as most people that I know edit, crop, convert, or resize picture files when they export them from their camera.
    • In particular, the instant that I started to read the NYT story, I dope-slapped myself for not having thought of the reverse implication of the technology, namely that it might be used to prove that a contraband image (such as child-porn) is NOT faked (and therefore is genuinely illicit).

      That's a good point. Hmm. Perhaps a free plugin for Photoshop/GIMP could be released and widely distributed that modifies an image to be in conformance with the model? That'd retain plausible deniability.

      My first thou
  • ******
    "With today's technology, it's not easy to look at an image these days and decide if it's real or not," says Farid. "We look, however, at the underlying code of the image for clues of tampering."
    - Hany Farid

    Farid's algorithm looks for the evidence inevitably left behind after image tinkering. Statistical clues lurk in all digital images, and the ones that have been tampered with contain altered statistics.
    *****

    doesn't really say anything beyond "hey we just look at the matrix code you'll see irregul
  • by Slothy (17409) on Saturday July 24, 2004 @01:41PM (#9789596) Homepage
    Run it on the goatse guy! If the results are positive I can finally start sleeping again.
  • by walmass (67905) on Saturday July 24, 2004 @01:42PM (#9789605)
    There are stories [usatoday.com] about successful defense against digital photographs in criminal cases. "Enhancements" using photoshop can be considered evidence tampering. So this technique can have a life-altering implication for some people.
  • by flynt (248848) on Saturday July 24, 2004 @01:43PM (#9789607)
    Anyone who wants to know more can read the actual papers here: http://www.cs.dartmouth.edu/~farid/publications/ ...papers 1,2,4, and 5 seem to be what this research is about, but others probably cover it too.
  • by Billobob (532161)
    for those who care about it...

    For Doctored Photos, a New Flavor of Digital Truth Serum By NOAH SHACHTMAN

    Published: July 22, 2004

    From the material found on his hard drive, Bryan Sparks of Springfield Township, Ohio, seemed guilty when he was arrested in 2002. The sexually explicit pictures of minors appeared to put him on the wrong side of child pornography laws. But at his trial this spring, Mr. Sparks was acquitted because no one could tell for sure whether the images were authentic or just clever


  • GWB: Lets just creatify some photographs of those Iraqi WMD

    DR: Great plan Mr President

    GWB : Dang! It won't work, they'll provify its a fake

    DR: Not till after the election George... not till after the election.
  • by rpiquepa (644694) on Saturday July 24, 2004 @01:49PM (#9789650) Homepage
    I wrote about this technology a while ago, in "True or False? Investigating Digital Images [weblogs.com]." A keypoint is that the Dartmouth College team thinks that their technology, or a similar one, will soon be incorporated in the U.S. legal system to authenticate images. At the above link to my blog, you'll also find an analysis of a forged image and more references, including the full research paper published by the IEEE Transactions on Signal Processing journal.
  • by randyest (589159) on Saturday July 24, 2004 @01:55PM (#9789679) Homepage
    Farid and his students have built a statistical model that captures the mathematical regularities inherent in natural images. Because these statistics fundamentally change when images are altered, the model can be used to detect digital tampering.

    That's pretty much all the detail on the method to detect image altering. Seems reasonable, but:

    1) How many real photos deviate how far from the statistical "norm" (i.e., how likely are false positives when checking for alteration?)

    2) How long before there are tools that can inject the proper (expected) statistical characteristics into a faked image?

    These are not addressed in the article. Anyone have more info?
    • The intro is not accurate. Actually, they are detecting regularities in the forged pictures.

      I've looked at their scientific paper [dartmouth.edu] and their technique albeit not perfect seems to be very good in detecting any kind of resampling in the image (up- and down scaling, rotating, etc.). When you make the transformation on a grid, the interpolation creates some almost invisible artifacts and regular patters which they are able to find by their analysis. It's difficult to create forgieries without these kinds of ma

  • reverse-engineering? (Score:2, Interesting)

    by chachob (746500)
    perhaps the technique could be reverse-engineered to allow the forgers to know whether or not their images can be detected as forgeries, and use this information to enhance their forging techniques to evade the detection tools...
  • Using the fake picture scanner backwards gives you a proofing tool to make sure you've perfected your fake picture before it's released. It won't take too long before these detection tools get on to eMule - and once there there, everyone will be able to fine tune their picture until the proofing tool no longer thinks it's a fake... Once in this state - they can release it knowing that the experts will be on *their* side. So such tools will have little use... It kind of works in the same way that you can us
  • by sameb (532621) on Saturday July 24, 2004 @02:03PM (#9789711) Homepage
    If there is an algorithm that can detect whether or not a photograph is forged, wouldn't it be possible to have another 'competing' algorithm that randomly altered bits and used the detection-algorithm as its success scale? (The better it becomes at not being detected, the better the genetic algorithm is doing.) Seems that that'd work fairly well, and would apply to lots of other technologies that require seamless photograph overlays.
  • "Suppose Nikon e.g., were to bury a private key in their cameras and use it to sign the raw image. Then the corresponding public key could be used to verify the image."

    Actually, I don't think that'd work, but it raises a very interesting point:
    when an UN-faked image is published with any kind of intentional alteration for legitimate purposes -- e.g., digital watermarking to protect the creator's copyright -- could that alteration have the unintended side-effect of making it impossible to prove that the ima
  • Details can be found in a preprint of the paper here [dartmouth.edu].

    ("RTFP? I can't be bothered to RTFA!")
  • You don't stretch images; you condense them. The source material must always be higher resolution than the output. Final pass is a very slight Gaussian blur, then an image resize down to the desired scale. This guy is detecting stretched images, which isn't hard.

    Or, as artists say, "work big".

  • There have been a lot of comments so far saying that this algorhythm would be a great help in creating doctored photos. Someone even said that if it passes the algorhythm surly it will pass the human eye.

    This may not be the case. There are certainly some features that we are very good at recognizing when there is something wrong: like proportions and white balance changes. I mean we are very good at this so creating an algorhythm to do better than us would be very hard. However there are some feature
  • A lot of people have posted about ways to defeat Farid's detection algorithm.

    1. I'd never say, "Gee, he's a professor, so let's all just trust him, because he must know what he's doing, right?" But people in his position do tend to be somewhat conservative about pre-publicity, to avoid later looking foolish and damaging their rep. In any case, that's why there's a peer-review process -- which is why I'd put a lot more faith in statements from him, than I would if they came directly from DHS.

    2. In the
  • from the NYT article- "for now the technique does not work as well with files created in JPEG ...At 90 percent quality, it falls apart very quickly," Professor Farid noted.

    doesn't this make the technique almost useless?

  • Detect those areas that show as fake, and perturb them until they don't. Iterate until the image passes the "authenticity" test.

    If you can identify why something looks fake, that information should imply a solution for non-fakeness.
  • To preface, I don't belive this at all. Consider a photograph that has undergone a hue transformation, and is titled George W. Bush has blue face, see photo. It will have all the other properties of a real photo, but will be fake. Another example, consider a photograph of models.

    But assuming this is possible, then I can make a fake photograph that is so convincing that they will belive it is true. First construct the photograph, and possibly include markings as to what are the edges of adjusted areas, etc

  • In one word: bullshit.

    Just like the "face recognition systems" used to spot terrorists [cio.com], or the "child protection" software that's supposed to recognise porn [dansdata.com].

    Not only is the success rate well below 90%, but, more importantly, it spits out thousands of false positives.

    And this is without even considering their admission that their technique does not work on JPEG images, even at "90% quality". In other words, they admit it won't work at all in 99% of digital pictures.

    These are all tasks that need so much
  • Is this photograph [ernet.in] real or faked?

    Bonus points if you answer the question without using the technique described in the article.

    P.S I'm not claiming credit for it; it used to be on his website but isn't anymore.

    • Yes, even to someone who didn't know the parties involved, your linked JPG is obviously fake. But I don't understand your point, unless it was humor.
  • One good way to authenticate an image would be to embed a GPS with a camera. Seeing as this is done on relatively inexpensive cell phones, it shouldn't be too hard. Market the camera as "secure" or "evidence-class" camera, somehow tie the info to the image...

    There are security details, but if it is possible to tag an image with WHERE it was taken in a secure fashion, it would eliminate all but the most enterprising forgers.
  • in the courts. Why? Simple.

    Our hope, however, is that as more authentication tools are developed it will become increasingly more difficult to create convincing digital forgeries

    Whos to say if a certain method is the right method to use? Just like there are numerous viruses, anti-viruses, and anti-anti-viruses; I think its safe to say there will photo manipulation software, counter-photo manipulation software, and counter-counter-photo manipulation software. Course this leads to programs being compromised

  • from the author (Score:3, Informative)

    by Hany Farid (799807) on Saturday July 24, 2004 @03:28PM (#9790192)
    In response to some of the posts on our work, we have developed a number of different techniques for detecting certain types of tampering. For those interested in technical details, please see:

    sp04.html [dartmouth.edu]
    ih04.html [dartmouth.edu]
    sacv03.html [dartmouth.edu]

    And, we have two new papers currently in review (abstracts are currently on-line, and preprints will be available soon):

    sp05a.html [dartmouth.edu]
    sp05b.html [dartmouth.edu]

    Some of these techniques work, as some have pointed out, only on high-quality jpeg or uncompressed images, while others work on lower-quality images. We are only in the early stages of development, and are currently working to extend some of these ideas to low-quality jpeg and gif images (though this will likely be a harder problem given that the compression artifacts will overwhelm any statistical perturbation resulting from tampering). One outcome of this may be that a legal standard is set that enforces images brought into a court of law to be of a certain resolution and compression quality.

    I will be the first to admit that each of the techniques that we have developed can be reverse-engineered, though doing so is more difficult for some techniques than others. It is our hope, however, that as we and others continue to develop more techniques it will become increasingly more difficult (though never impossible) to simultaneously foil each of the detection tools.

  • by dpbsmith (263124) on Saturday July 24, 2004 @06:30PM (#9791050) Homepage
    The New York Times article actually says something about the methods used. It basically works by looking at the kinds of operations that need to be performed when interpolating between pixels in manipulations such as changes in size, rotation, etc.

    And, "Professor Farid said that for now the technique does not work as well with files created in JPEG, the compressed picture format most commonly used online. As the size of a JPEG file shrinks, the correlations between pixels become much less obvious. 'At 90 percent quality, it falls apart very quickly," Professor Farid noted.'"

From Sharp minds come... pointed heads. -- Bryan Sparrowhawk

Working...