27

Suppose I make a photo and publish it and then find that someone else published it and claimed it's their photo. Or any similar situation - it doesn't matter. So one of the parties of such an argument decides to prove that they were indeed the author of the photo and the other party just unfairly claims authorship.

I heard owning a raw version of the photo is a commonly accepted proof. The author somehow shows that they have a raw version and the other party doesn't have such and so the other party loses the argument. This assumes the author doesn't initially publish the raw photo of course.

What if the photo was taken in JPEG initially? There're many cameras which don't output raw at all - they only output JPEG. Now one person has a JPEG and another person has a JPEG and how do we know who was the author?

sharptooth
  • 1,059
  • 1
  • 13
  • 19
  • 5
    I can't help but wonder... is this a hypothetical question, or are you actually facing this issue for real? – osullic Aug 07 '18 at 11:58
  • 2
    This question is inspired by this story https://pikabu.ru/story/_6075720 in Russian - A claims that B submitted A's photo to a contest under B's name and... won the contest under B's name. I've also read a number of stories where photos are used without permit and sometimes that even turns into a lawsuit. – sharptooth Aug 07 '18 at 12:28
  • 2
    And so photo.se was flooded with questions about SWIM – xiota Aug 07 '18 at 18:28
  • There is the possibility of the camera leaving a "signature" in the image. This can be dead/hot pixels, or more complex things which may or may not be accepted by a court – PlasmaHH Aug 07 '18 at 19:03
  • Jurisdiction might matter for details on this question. For example, in the US, generally copyright infringement suits are thrown out if the plaintiff has not registered the copyright with the Library of Congress. Registration generally establishes authorship in the legal sense. So if you want to protect your JPEGs, register them. – Todd Wilcox Aug 07 '18 at 20:19
  • 3
    @ToddWilcox, you can't file suit if you haven't registered the copyright, but you can register the copyright after the infringement and still sue. The only restriction that comes from not registering is that, for infringement prior to registration, you can only sue for actual damages, not punitive damages. – Mark Aug 07 '18 at 22:05
  • 1
    @Mark As long as the copyright is registered within 90 days of first publication, one may still also file for punitive damages (unless the changes effective February 20, 2018 changed that), even if the infringement preceded registration. No one mentioned such a change in the summaries I saw, which mainly dealt with how many images may be included in a single filing and how "single author" is defined for groups of images done by different individuals as work for hire for the same person/entity. – Michael C Aug 08 '18 at 08:05
  • Interesting paper: https://arxiv.org/pdf/1707.07795.pdf Anti-Forensics of Camera Identification and the Triangle Test by Improved Fingerprint-Copy Attack – Mark Plotnick Aug 09 '18 at 17:13

7 Answers7

45

Throw away the technology for a second, and consider before digital. Before digital, the negative was defacto proof: there was (typically) only one, and the author had it. But if there was no negative, due to loss or damage, then standard detective/police work is needed: Proof that the photographer was in the location when the shot was taken; testimony of others in the shot or at the location, other photos taken in the same location at same time, etc.

The situation is the same here: you would need additional evidence that the author was at the location at the time of the shot: mobile phone records, GPS, and witnesses. The easiest item is to produce additional photos at the same time and location: the imposter would not have other images.

cmason
  • 14,792
  • 1
  • 34
  • 58
  • 3
    It is no problem to make a copy of a negative. When done correctly, the copy will likely be so good, that it is impossible to distinguish it from the original. – jarnbjo Aug 07 '18 at 12:31
  • 8
    @jarnbjo - Yes, but I'm guessing that, much as a digital photographer is far more likely to share jpegs than raw, a film photographer is far more likely to share prints than negatives. Hence "there was (typically) only one". – AndyT Aug 07 '18 at 13:33
  • 4
    If both the original negative and copy are available, I would expect an expert/scientist to be able to distinguish which was which. – db9dreamer Aug 07 '18 at 16:09
  • @AndyT If the print is of sufficiently high quality, you can also make a high quality negative copy from a print. You don't need access to the original negative. – jarnbjo Aug 07 '18 at 16:58
  • 7
    @jarnbjo The copy negative can never contain information that an original does not. The obverse is not true. It is almost always possible to forensically determine which of two negatives is the one from which the other was sourced, either directly or indirectly via an intermediate print. – Michael C Aug 07 '18 at 19:48
  • 2
    Film cameras almost always left a pattern scratch marks on the negative that were characteristic of the particular camera, even if the marks were too small to regrade the print quality. If it's admissible legal evidence to match a bullet to the gun it was shot from, the same applies to film and cameras. A copy of the negative would not preserve those "fingerprints". – alephzero Aug 07 '18 at 23:12
  • 1
    @alephzero: ...unless the copy was shot in the same camera as the original, of course. – Jerry Coffin Aug 08 '18 at 00:18
  • @jarnbjo I would wager that if somebody's stolen your negative from your vault/safe/lab, you have larger problems than someone claiming you are misattributing their work. – Lightness Races in Orbit Aug 08 '18 at 10:57
  • @LightnessRacesinOrbit Noone has to steal anything to get access to my negatives. It would be enough that I give them to a lab for printing, which is quite common. – jarnbjo Aug 08 '18 at 11:24
  • @jarnbjo Then the lab has to steal them – Lightness Races in Orbit Aug 08 '18 at 11:43
  • @LightnessRacesinOrbit Steal what? A lab can obviously make a copy of my orginial negative and return the original negative to me without me noticing. – jarnbjo Aug 08 '18 at 11:45
  • @jarnbjo Which is stealing :D My point is that if the lab you use is a point of failure to that degree then the specific photo stolen is not the biggest problem you have. I'm not suggesting that this actually helps you but it does reframe the task at hand – Lightness Races in Orbit Aug 08 '18 at 12:54
  • @jarnbjo Just because there are a few, very detailed (and wildly speculative) scenarios in which no forensic analysis can conclude which claimant took an image does not logically require that all such cases can not be resolved using forensic methods. Very few criminals think of every conceivable way to cover their tracks. That's often how they get caught. Let's say your lab employee steals your negative and returns a duplicate to you. But the date and time he claims he took the photo doesn't match the weather, sun position, or other details in the scene. It is later shown that he was... – Michael C Aug 09 '18 at 02:40
  • ... conclusively documented to have been somewhere else at the only time/place the image could have been made, based on the contents of the image. – Michael C Aug 09 '18 at 02:43
28

You can prove that the picture is yours if you have other data that are not in the subject picture:

  • the RAW file from which the JPEG can be produced (framing/perspective)
  • a higher-definition picture from which the subject JPEG has been scaled down
  • a larger picture (at the same definition) from which the subject JPEG has been cropped or some details removed (for instance, that phone pole which is there IRL)
  • a series of pictures taken at the same time and place from which the subject picture is extracted
  • all of the above :)

Basically, never post all the pixels of the shot...

xenoid
  • 21,297
  • 1
  • 28
  • 62
  • Good points, except for your first one which is explicitly ruled out by the OP. – Robin Aug 07 '18 at 14:54
  • 12
    Yes, but I wanted the answer to be complete. – xenoid Aug 07 '18 at 17:29
  • 2
    Not sure about the conclusion: arguably, the cropped picture may (should?) be better than the uncropped one – clabacchio Aug 08 '18 at 07:24
  • @clabacchio But if one party can produced an uncropped image and the other party can only produce the cropped version that was published by the first party... – Michael C Aug 09 '18 at 02:45
10

There are several possibilities, and the hierarchy will vary from one place to the next.

But in general, the idea of precedence is the starting point. That is, who published the image first? Absent of any other compelling evidence to the contrary, the person who first published the image will probably be recognized as the author of the image.

Things that could possibly override the earliest original publication date:

  • If one party can produce a higher resolution version of the image than the other (that does not appear to only be an up-scaled version of the information contained in the smaller sized version), or any other more pristine version of the image. For instance, a less compressed version of the image at the same resolution would carry similar weight as a higher resolution version would.
  • Copyright information in the metadata of both images that agrees that the same person is the author of both images. Almost all major publications require the IPTC metadata in an image to include copyright owner/author information.¹
  • A camera or lens serial number in the metadata of both published images that matches a camera or lens to which one but not the other had access.¹
  • Any other details in both copies of the image or its metadata that can establish a specific camera produced the image if only one of the two persons claiming to have the produced the image had access to that specific camera.²
  • Corroborating evidence that one but not the other person was present at the time and place the image was captured. If the image is from a specific event and one person was issued a media pass that placed them in the position to have captured the shot from a "press only" area while the other person can not demonstrate they were present at the event then the person who can demonstrate that they were there probably prevails. This would be particularly the case if everyone who had access to the place from which the image was recorded required documentation/registration/media pass/etc.
  • The testimony/deposition of any persons pictured in the photo, or person(s) who witnessed the photo being taken, as to who took the photo.

¹ Please notice that all references to metadata as valid for determining authorship is placed in the context of both versions of the image as published by both parties having the same metadata content for the particular fields in question. It would also be based upon the premise that the actual image information can be shown to have been produced by the camera indicated in the metadata. If the metadata does not agree, then it would behoove both parties to make their case based on other factors, or to establish that their version of the metadata is the original version using whatever forensic methods they may have at their disposal. Please see note 2 below for more regarding congruence between image information and metadata.
The question seems to demonstrate no awareness that such information can even exist within the image file itself. It therefore seems possible that a party who violates someone else's copyright might also be unaware of such metadata and publish the image as their own without altering the metadata. In such a case, establishment of ownership would likely be resolved long before going to trial.
² For instance, if the content of the actual image data does not match those used by the camera indicated in the metadata, that would raise a red flag that the metadata has, in fact, been altered. It could be one or more of any number of things: Does the compression algorithm used, or the pixel mapping used to map out hot pixels, or other sensor characteristics revealed by the actual image data, match the camera referenced in the camera information fields of the metadata? Is the serial number in the metadata a valid serial number for the type of camera that took the image? Can the camera indicated in the metadata have been used to produce the image in question, or is there some characteristic of the actual image contents that eliminates any possibly doctored metadata, such as camera identifying EXIF fields, from being valid?

Michael C
  • 175,039
  • 10
  • 209
  • 561
  • 4
    I sincerely hope that image metadata is not considered proof of anything by a court. Any child can take one of your image files and insert metadata with my equipment's serial numbers. – jarnbjo Aug 07 '18 at 10:53
  • 1
    @jarnbjo I have absolutely no clue about legal things, but I suppose it would be up to your legal representative to explain that to the court. – osullic Aug 07 '18 at 11:57
  • @jarnbjo first of all, the internet is not the place for anyone to obtain legal "definitives". Now, isn't it the case that courts listen to evidence and make judgements? So, legal representative A uses metadata as "evidence" of something. Legal representative B refutes that metadata proves anything. That's how I see it working. I'm certainly not arguing with you - as I plainly said, I have no expertise in this area - but I will certainly repeat that the internet is not the place for anyone to be getting answers to questions of law. – osullic Aug 07 '18 at 14:23
  • @osullic If the court is not able to assess what is presented as alleged proof, it will call an expert witness and not blindly believe what a lawyer claims. The expert witness will then tell the judge that metadata in a digital image file has absolutely no value when the question is from where the file originated. – jarnbjo Aug 07 '18 at 15:13
  • @jarnbjo, I'm sure that it would be very amusing to listen to a defence lawyer explain how a child had edited the metadata of their client's photo to insert the valid serial number of someone else's camera before their client uploaded the photo to their own website. But you may be alone in believing that it would convince the judge or jury. – Peter Taylor Aug 07 '18 at 15:41
  • @PeterTaylor That is not what I wrote. Please don't invent things. I can take an arbitrary image file and add traces of my own equipment to the metadata. Nothing more, nothing less and it proves absolutely nothing. – jarnbjo Aug 07 '18 at 15:58
  • 3
    @jarnbjo, so in that case why post it as a comment to this answer, which specifically and repeatedly talks about commonalities between the metadata of both images? – Peter Taylor Aug 07 '18 at 16:00
  • @PeterTaylor Because it can not generally be ruled out that whoever made the copy also had access to and was able to manipulate the real original version of the file. – jarnbjo Aug 07 '18 at 16:18
  • Please notice that all references to metadata in the answer as valid for determining authorship is placed in the context of both versions of the image having the same metadata content. If the metadata does not agree, then it would behoove both parties to make their case based on other factors, or to establish that their version of the metadata is the original version using whatever forensic methods they may have at their disposal. – Michael C Aug 07 '18 at 19:38
  • @jarnbjo "Because it can not generally be ruled out that whoever made the copy also had access to and was able to manipulate the real original version of the file." Generally, once an image has been published, the "cat is out of the bag." Any post-publication attempt to alter metadata of all extant copies of an image will prove to be extremely difficult. – Michael C Aug 07 '18 at 19:44
  • @Jarnbjo Rather hypocritical for you to accuse others of inventing things, when you turned "A lawyer would present an argument" to "A lawyer would dictate to the court what they will accept as proof". Witnesses are generally called by lawyers, not the court sua sponte. If a party were to present metadata in court, the other party's lawyer like would tell the court that metadata can be changed, and then the burden of proof would be on the first party to present evidence regarding the reliability of metadata. – Acccumulation Aug 07 '18 at 19:55
  • @Acccumulation Or the other party's representation would call an expert witness to demonstrate to the court how metadata could be changed. Or the other party's lawyer would cite existing case law that has already well established such. – Michael C Aug 07 '18 at 19:58
  • @Acccumulation Where I live, expert witnesses are often called by the court to interpret or evaluate evidence presented by a part. I am contradicting what osullic wrote in his comment. Lawyers will not be competent to substantiate the technical validity of proof in a court and will probably be clever enough not even trying to do so. It may all be different in the US fun fair court system though. – jarnbjo Aug 07 '18 at 20:28
  • @MichaelClark It is not difficult to imagine a potential situation where a crook will have access to and opportunity to modify an original image file before publication. I frequently order prints on self-service terminals in photo labs directly from my camera's memory card. It would be a no-brainer for the operator of such a terminal to both make a copy of the original raw file and modify my file on the memory card to include meta-data pointing to his equipment. Voila, the crook has in your opinion proof of being the proper author of the image. – jarnbjo Aug 07 '18 at 20:33
  • @jarnbjo Perhaps it might be difficult to imagine that no other scenario but that could exist, either. The word possibly is included in the answer for a reason. The word always was not used anywhere. – Michael C Aug 07 '18 at 23:01
  • 4
    For those who are overly concerned about the metadata, while it is easy to alter (exiftool -SerialNumber="0000" image.jpg), you need multiple lines of evidence to prove something. You cannot hang your hat on just one thing. That is why the answer currently has 4 paragraphs and 6 bullet points. All available information has to be considered. – xiota Aug 08 '18 at 08:15
  • @jarnbjo If the technician changed the metadata on the version of the memory card that was given back to you and you tried to view the image in the camera in which it was taken, it very likely would not display properly. Very few cameras will display images that have been altered by a second device. None of Canon's cameras that I'm aware of will display an image that has been altered in any way. They will display unaltered images taken with other Canon cameras, but if the checksums do not match they're rejected as corrupted. This is probably done as a security feature to prevent viruses. – Michael C Nov 02 '21 at 09:43
6

It might also be worth noting that cameras have a specific signatures in the form of the noise pattern which apparently is worth, here is one article talking about this:

Hidden “Signature” in Online Photos Could Help Nab Child Abusers
A new technique exploits sensor noise patterns unique to each camera that can help identify criminals via photographs posted online

The article talks about how this could be used for forensics, but maybe for this purpose here as well, especially if we are talking about multiple photos.

brasofilo
  • 195
  • 1
  • 10
phk
  • 161
  • 4
3

Steganography (messages hidden in plain sight) can prevent this situation

A steganographic watermark can be used on digital images to hide a message within the image. This would allow an author to put a secret ownership message within the image itself without altering the image's colors enough to be seen by the naked eye.

enter image description here (https://www.endgame.com/blog/technical-blog/instegogram-leveraging-instagram-c2-image-steganography)

Even if a person were to screenshot the image from the web or download it and spoof the EXIF data, the secret message can be extracted. You can place in a message into the image saying "Copyright Sharptooth" in random spots. If there is ever question to who owns the image, you can extract the steganographic message and reveal that you are the owner.

For more information on how image Steganography works, check out this video by Computerphile: https://www.youtube.com/watch?v=TWEXCYQKyDc

Lil' Bits
  • 131
  • 3
  • 3
    Ordinary people call this digital watermarking. People being watched by certain government agencies call it stegosaurusy. Also, depending on methods used, it's unlikely that it would survive through even JPEG rounding errors intact. – xiota Aug 08 '18 at 01:56
  • 1
    @xiota: if someone recompresses the image, it will necessarily lose at least a little quality. That's normally detectable (see other answer), so you can either prove that you have a higher-quality copy of the image than the other person making a claim, or their copy is bit-exact or high-enough quality to preserve the steganography. Hiding a small message over a whole image can (I think) be done fairly robustly, in ways that will survive recompression at any nearly-transparent quality level. – Peter Cordes Aug 08 '18 at 04:11
  • 2
    @PeterCordes My comment is specifically about stegasaurus-watermarking. The naive approach is to modify the least significant bits of an image. This would not survive JPEG compression, even at 100%. A JPEG-specific approach, such as modifying the quantization coefficients, would also not survive JPEG-recompression, and it certainly wouldn't survive a screenshot. Stegasauruses aren't particularly robust. That's why they're extinct. – xiota Aug 08 '18 at 05:41
  • 1
    @xiota A good watermarking algorithm is designed to withstand compression. Hiding the message in the LSB is not good watermarking. It is true however that, as always in information security, the method depends on your threat model. – Alex bGoode Aug 08 '18 at 10:34
  • 1
    Steganography refers to messages hidden in the medium with the purpose of concealing their presence. What you're talking about is indeed called watermarking. Of course you could use steganography for the purpose of watermarking, but it would be suboptimal. – Dmitry Grigoryev Aug 08 '18 at 11:43
  • I'd add or subtract insignificant random amount of color to every pixel - noise kills steganography, and if noise level is small it is not easy to say whether a noise was added to clean picture or a clean picture was de-noised. – Barafu Albino Aug 08 '18 at 21:56
0

Time stamps of each image would be one way. Obviously, if you posted your image first, it's timestamp would be earlier than the stolen copy.

EXIF data. In many cameras, you have the ability to add copyright information right into the EXIF data. Plus, if you have the original, then your image will have all of the EXIF data. Images uploaded to the web are commonly stripped of their EXIF data in whole or in part. Some cameras will write the model & serial number of the image. GPS information could allow one to prove that the alleged perpetrator couldn't have taken the shot since they were never in the specified location.

Image Size: If you upload lower-resolution images, then only you will have the full-resolution image while the person who uploaded the stolen copy can't.

There are apps that will embed copyright data right into the image itself which is impossible to remove.

Frank
  • 674
  • 3
  • 7
  • 2
    Timestamp where? If the timestamp is part of any data hosted by your website, then you have control over it, and so it is not reliable. I suppose if your website is archived by another website, then that website's timestamp would be admissible evidence. – Acccumulation Aug 07 '18 at 19:57
  • GPS isn't useful here. If one of the contenders has really taken the picture, s/he knows where this was, and can prove it with shots from other folks (or Google's StreetView....) if it was taken outdoors or in a public building. – xenoid Aug 07 '18 at 23:50
  • 1
    AFAIK EXIF doesn't have a proper digital signature of the photo, so I could take a photo similar to yours, rip the EXIF from it and attach to your photo. – Dmitry Grigoryev Aug 08 '18 at 11:31
  • Accumulation - The timestamp is embedded within the file's metadata. – Frank Aug 08 '18 at 21:29
  • GPS can be useful. For example, the photo was taken in a specific area of Idaho, but the offender cannot prove that they've ever been to Idaho. – Frank Aug 08 '18 at 21:31
  • 1
    @DmitryGrigoryev That is correct. But you don't even need to "rip" something from anything, because EXIF data is trivial to edit. There is lots of software to do that and it doesn't leave any traces in the file. So EXIF data does not prove anything. I would downvote this answer, but I don't have the necessary reputation on this site. – Philipp Aug 09 '18 at 08:28
0

Having the original in JPEG is not that different from having it in RAW.

If the author published the original JPEG straight from the camera, they will need proof besides the original image to prove authorship. Typically the author would have other shots of the subject/event that they didn't publish. It's also possible to prove that the shot was taken by a particular camera (by analysing the sensor defects which result in similar artefacts in other photos). If the author owned the camera, they would have no problem presenting other shots they have done with it.

If the author processed the original shot before publishing, they could prove authorship by releasing the original shot together with instructions on how to process it. Generating a fake unprocessed shot in JPEG isn't any easier than generating a fake unprocessed shot in RAW.

Dmitry Grigoryev
  • 316
  • 3
  • 10