Slashdot Mirror


Digital Camera Image Verification

Polo writes "While reading dpreview, I noticed that among several new products, Canon has announced a Digital Image Verification Kit to prove that an image taken by a particular camera has not been modified. It's disturbing to think about the conditions that would allow digital images to be accepted in a courtroom. I guess one defense would be to figure out how to 'verify' a photo of shark attack..."

7 of 255 comments (clear)

  1. Re:It's called MD5 (?) by filtersweep · · Score: 4, Informative

    "All in all I suppose it's a neat idea -- hope it actually works before somebody is on trial for his life though..."

    Well, the camera is only one step in the chain. Are they going to keep a bunch of these presumably more expensive memory cards lying around, or are "they" going to archive them on a CDR or hard drive? Once the image is out of the card, the verification is meaningless (if it wasn't already meaningless in the first place).

    I provide "expert testimony" in court on a semi-regular basis in a completely different field. I always submit "photostatic replicas" of original documents and sign a notarized affidavit of their authenticity. Overall, it is simply the sworn testimony of the authenticity of any evidence that holds more weight than some "technological solution."

    Photoshoppers be dammed! Long live fark.com

    --


    Those that suggest you "dance like no one is watching" really want to see you make a complete fool of yourself.
  2. Re:won't work by contrasutra · · Score: 5, Informative

    Haven't read the gnupg.org website? From the front page:

    GnuPG's ElGamal signing keys compromised (2003-11-27)
    A severe problem with ElGamal sign+encrypt keys has been found. This leads to a full compromise of the private key.


  3. Still does not by www.sorehands.com · · Score: 4, Informative
    Even when taking a photo, to have it admitted as evidence you must have the person taking it verify that they did take it. This goes with digital or film camera -- or any type of documentary evidence.

    This is just general, but there are many rules about entering photograghs and other documents.

  4. Re:2D autocorrelation... by Rosco+P.+Coltrane · · Score: 4, Informative

    I've been wrestling with the idea of writing an image modification detector

    Forget it. Only amateurs copy/paste regions and leave them like that. Those who alter images to produce really credible results may copy/paste bits of images at first, but then will blur/sharpen/solarize/burn/lighten/brush slightly part of them, drop some noise in them to match the pizelization of an original jpeg for example, merge several together and modify gradiants to make the final patch blend in just right in the bit of background you want to mask or change. The final resulting altered regions usually doesn't have much to do with the original bits you copied.

    --
    "A door is what a dog is perpetually on the wrong side of" - Ogden Nash
  5. Re:Courtroom.Rules of evidence by Anonymous Coward · · Score: 3, Informative

    This is correct. The federal rules of evidence (and the rules of most states) require that the witness testify that the photograph actually depicts what it is that the witness says it depicts. The witness could paint the photograph, if he were an adequate artist.
    All writings and papers and so forth have to be introduced in such a way as to either not be hearsay or to gain a hearsay exception.
    I don't know why you might think that a video movie is more sacrosanct than something like a blood sample. Both require someone to testify about them and in both cases the person can convict someone simply by lying.
    By the way---remember the video in the Microsoft trial? They could have easily faked that too.
    Sooner or later you have to rely upon people to tell the
    truth and there is no way around that fact. These cameras will make no difference whatsoever.

  6. EXIF, distortions by wotevah · · Score: 4, Informative

    The camera stores information about focus distance, focal length (zoom) and exposure parameters as well as other data in each image (in EXIF format, commonly). Example:

    Camera make : SANYO Electric Co.,Ltd
    Camera model : J1
    Date/Time : 2004:01:15 14:21:22
    Resolution : 2048 x 1536
    Flash used : No
    Focal length : 6.0mm
    Exposure time: 0.400 s (1/2)
    Aperture : f/2.9
    ISO equiv. : 113
    Metering Mode: center weight
    Exposure : program (auto)
    (focus distance is manufacturer-dependent and jhead couldn't get it).

    Also, you'd also have to account for the distortion effects that are measurable and reproducible with each camera model. For example, barrel or pincushion distortions compound if you take a shot of an existing picture.

  7. Re:2D autocorrelation... by wotevah · · Score: 3, Informative

    Since you said "uniform noiseless areas (sky)" - funny thing is, the sky is one of the most difficult things to get an "uniform" picture of. All digital cameras I know of produce "sky noise" in various proportions.

    A picture of the sky is how you can quickly check how noisy of an image the camera can make (part of it can be internal image processing, of course).