The fingerprint inside photos
bbc.com
bbc.com
"Associating cameras with users and objects in a social networking system" https://patents.google.com/patent/US20150124107A1/en
Ok, but will that survive JPEG compression? Since JPEG is meant to remove details that aren't visible to the human eye it would be surprising if this fingerprint survived in a usable way.
> Technically, it should be possible to suppress, for example, by reducing the image resolution, says Farid. But, by how much? This of course depends on many factors such as the type of device used for image capture, as well as the fingerprint matching algorithm employed. There is no one-size-fits-all solution to fingerprint removal.
I'm skeptical this works after boiling 24...50 megapixels down into a tiny web-sized image with 1-3 MP.
Of course, in some cases shooting RAW from a specialised camera is better, but quite often JPEGs straight out of a standard (cheap DSLR or mirrorless, or webcam) camera are more than enough. Takes an order of magnitude less memory as well, which counts when you have stacks with 200 pictures.
There is a discussion here, for example: https://www.planetguide.net/astrophotography-webcam/
Errors in a CCD are not necessarily random - if the error is on a mask used in the silicon production process, you might find for example that hundreds of other sensors across the world all have a very similar error profile.
Knowing if your software has picked up two images taken by the same camera, or one of the few hundred made with the same masks is going to be awfully hard to do with sufficient confidence...
Yeah, AI may one day be able to fool the human visual system but it will never be able to add a watermark to a video.
Ultimately this is a classic case of security through obscurity, it won’t protect us from deep fakes long term.
They are apparently not doing anything like that, it’s just “hey, these photo sites are all a bit different”.
Well, that’s what they’re saying publicly anyway, could be complete bs and every sensor has a government sanctioned backdoor in it.
It's probably a good idea to use a common phone or cheap camera. And change date and time, and make sure it does not leak anything in the EXIF metadata. And make sure that the sensor does not have any dead pixel, because that would be a dead giveaway.
Maybe they should invert the 'quality' scale and call it 'privacy' ;-)
I like how they call the modern world. Though, I'm not sure what time in history can be called "truth age".
I guess the word "post-truth" means the lack of truth as an objective or goal, not the lack of truth by itself.
I'd say most of it, and we're still largely there.
People basically believe what they're being told and shown. Being inhertinaly skeptical is very rare.
This sad term makes me feel like there are adults out there still grappling with the fact that lies and uncertainty exist
I don’t actually see any of those as being difficult to implement - but I guess there’s also no market.
How could they? They have no physical shutter.
Tying a specific camera to one or a few photos in a criminal case sounds extremely dubious. Natural images are full of different kinds of noise anyway, and each pixel only has 256 levels of intensity.
I remember spending days/weeks 18 hours a day cold calling all my contacts one after the other because I wanted to "hear someone say they are alive".
In the end everything worked out pretty fine and a major disaster didnt happen.
(original seems to be down rn, here's the archive link): https://web.archive.org/web/20101229153120/http://www.hacker...
And HN discussions, from 2009: https://news.ycombinator.com/item?id=1017726
and from 9 months ago: https://news.ycombinator.com/item?id=23538059
I don't see why? Seems like all you need to do is take a photo of an evenly lit gray card, determine the variation of each pixel, and then apply the reverse of that to every photo you take.
And a simple filter can even get around the "evenly lit" part trivially, as long as it's smoothly lit. (I.e. handle gradients.)
The article doesn't mention this at all -- yet isn't this the kind of thing astronomy photographers do when stacking photos, to prevent the "fingerprint" from accumulating?
Of course, it's not even clear this "fingerprint" is even remotely readable from average photos -- it seems like it would be completely indistinguishable in any scene with detailed textures. So it's really only in areas of blank walls, sky, etc. where it poses a risk.
Astronomers (I presume) do much more in the way of non-uniformity correction (NUC). Often one of the biggest differences between consumer cameras and industrial/scientific ones is the ability to continuously recalibrate them, either in firmware or in post. You could do the same with off-the-shelf cameras and some software, but there's little need most of the time.
Can you explain why? Is it not just a simple subtraction?
E.g. is it not a pattern of predictably varying intensity per-pixel, but a spatial pattern of nondeterministic noise? So each individual pixel has a likelihood of contributing to the "fingerprint" in any given photo, but you don't know if it contributed in this one?
Or does it totally predictable on the sensor, but the on-camera RAW processing followed by JPEG compression transform it in non-predictable ways, so that you'd need to extricate it on RAW data directly before any processing?
Or is it something extremely non-linear that changes depending on pixel brightness per color channel, or is drastically different depending on exposure time?
I'm super curious if you or anyone has actual knowledge here.
There is a reason why 3D rendering is difficult; light bounces an awful lot. Objects slightly tint the whole room even if unperceptibly. Lamps vary in intensity, and their shape affect the outcome. So how do you even set the scene up? With a small card the phone being so close can affect lighting. A distant large surface requires more distant lights for uniformity, o simply more lights. Getting the camera out of focus might be useful, but can't correct large scale phenomena properly. Getting the watermark is the difficult part imo.
After all that, once you substract it from every picture you take, every subtle mistake you made in the watermark generation phase will become your new watermark. You bet your privacy on your error being very close to the zero function across the whole grey card. As mentioned in every other comment, astronomers do these calibrations often... for a living. I'm simply not that invested beyond removing EXIF data.
Perhaps I'm stupid but could it be possible that the camera clocks were not set correctly?
There could be several potential sources for this "timestamp," but I haven't seen the one they are talking about actually mentioned.
I believe this is done for handguns - they test fire the gun, then record the way the firing pin marks a cartridge - all before it is sold.
What if I have cropped and re-saved using GIMP or Mac's Preview?
That is all.
And regarding the timestamp: it isn't that accurate after all, because the timezone is 'local' (whatever that means). So if you don't know where the picture has been taken or can't be sure that the device actually uses the local timezone (who changes the time on his DSLR when he travels to a different timezone?), you can only derive the date with certainty.
After all, whenever you post an image online you should be aware of the information you put out there. For example, if you are an American and in the background of the picture is the Eiffel Tower, a burglar doesn't have to be Sherlock Homes to guess that you are on vacation.
In addition, the article is about the information that is not explicitly encoded into the image (like EXIF), but implicitly reveals the sensor fingerprint, which you would not see even if you would take a look at every byte in the file, because you need to know what exactly to look for. If you know what too look for, you could theoretically see it even on the printed version of the picture ;-)
It was too little, too late. Pros never took to it, and the smaller size and variable aspect made it less-than-ideal for consumer printing.
Almost all social media strips this data other than really amateur stuff or 4chan. Or in the case of vice, their own site.