7 ms·
Sony Unlocks In-Camera Forgery-Proof Technology
- Rackedup 4y agoWhat if you take a picture of a modified picture?
- jack_pp 4y agoI like this, if we can fool humans with VR goggles then surely we can fool the camera by using a sufficiently detailed print or a high resolution display
- CharlesW 4y agoA more effective version of this would capture a 3D depth map with the 2D image.
- ChrisLomont 4y agoYou'll never remove aliasing artifacts.
- TeeMassive 4y agoNot all fake images are images taken from cameras (e.g. CGI, AI)
- josephcsible 4y agoILM figured out how to: https://en.wikipedia.org/wiki/StageCraft https://en.wikipedia.org/wiki/StageCraft
- ChrisLomont 4y agoWhat in that makes you think there are no aliasing artifacts? Yes, you can take a movie of a movie, and you introduce artifacts that simple frequency analysis will find. There's no reason ILM would even try to remove this for such an application - they simply want pretty visuals, not mathematically indistinguishable from reality digital compositing. Take a photo. It's not a grid of values. You already screwed with high frequency components. Take a photo of that - you changed these again. It's not hard to detect this. There's an entire field for detecting forensically altered things, and no current tech goes undetected, certainly not stuff simply made for movies. Or provide some link, paper, or serious scientific claim that ILM has magically removed all aliasing from images, which would, well, simply violate physics. What they actually do is try to move aliasing to less visible portions of a frequency spectrum - but those signals are still there and detectable.
- praash 4y ago> -- would, well, simply violate physics. Nope. Aliasing can be eliminated by low-pass filtering (blurring) the input signal before sampling it. This could be done by keeping the background even slightly out of focus, or coating the display. Unless it's traded away for performance (or style choice), you are experiencing alias-free sound and image processing all day long.
- ChrisLomont 4y agoYep. >Aliasing can be eliminated by low-pass filtering (blurring) the input signal before sampling it. Nope. I have a PhD in math, and have been doing filtering stuff for decades. You do not eliminate aliasing, you push the error out of the visible spectrum - end of story. The naive, physically impossible undergrad textbooks might say differently, just like they'd teach Newton as gravity and ignore relativity, or teach the ideal gas law and ignore that real gasses are vastly more complex in behavior. This is Wikipedia level knowledge: from the low-pass filter page [1]: "However, the ideal filter is impossible to realize without also having signals of infinite extent in time, and so generally needs to be approximated for real ongoing signals, because the sinc function's support region extends to all past and future times." So no ideal filter, sorry. That means leakage, which happens in every part of engineering (and is forced by physics, despite you not wanting to believe it). Ideal filters are math, not physics. And this level is vastly below what is done in industry, what theory (and physics) forces onto real world filtering. Once you discretize the signal, even after filtering, you have introduce lots of noise in lots of frequency ranges. The simple reason is that you are forced to perturb real sample values (which would require infinitely precise measurements) to the discrete ones your sensor records. These perturbations do not fit your frequency requirements. Don't believe it? Take a photo, apply your "ideal" low pass filter, then fit the resulting image using sin and cos waves of frequency no higher than than your filter - and you'll see it's impossible. You can only approximate it - and that is empirical proof you can do at home. Another entire level that undergrad level understandings miss is that pretty much zero of the theorems you think apply actually do - because pretty much every one uses infinite precision samples (Nyquist, for example), infinite support reconstruction (physically impossible), and so on. In practice all of these things leave noise in various parts of frequency spectrums. So, still want to argue that filters would somehow remove aliasing issues? Especially over a sensor with the number of samples and precision that a modern Sony sensor has? >you are experiencing alias-free sound and image processing all day long You are confusing Nyquist with reality - and ignoring that Nyquist doesn't even apply to audio (or video) - it's only an approximation to what is really happening. As an approximation it's useful, but the requirements of the theorem are not met in reality and the reconstruction is not met in reality. Don't believe me? Go look at the proof for the Nyquist theorem, and pay attention to all the requirements that no physical device can meet. it's the frictionless spherical cow of audio processing. One glaring area it fails is the need for the infinite support sinc filter - which again is why real engineering is not so simplistic. [1] https://en.wikipedia.org/wiki/Low-pass_filter https://en.wikipedia.org/wiki/Low-pass_filter
- TeeMassive 4y agoThen you have a signed picture of a modified picture which can't be altered further. Cryptographic signatures only ensures data integrity and authenticity after the cryptographic signature was made.
- remram 4y agoIt can totally be altered further. Just print the altered version and take a picture again.
- TeeMassive 4y agoNot without having a valid signature
- remram 4y agoDo you mean "not with"? The camera generates a valid signature every time you take a picture...
- Rackedup 4y agoreminds me of NFTs... go Sony...
- josephcsible 4y agoWhat's stopping me from sending fake data to the signing chip along the same route that the image sensor uses to send real data to it?
- thaumasiotes 4y agoBecause, it's forgery-proof. Didn't you read the headline?
- karamanolev 4y agoThey either built some smarts into the sensor, where it's physically impossible to trick it or you can just pretend to be a sensor capturing whatever you want. I assume with enough desire, you can built a pixel-for-pixel matching LCD to illuminate each sensor photocell as desired and capture arbitrary light the physical way. But as another commenter pointed out "it's forgery-proof" ;)
- aaaaaaaaaaab 4y agoYou can't do that, because it's Not Allowed™.
- xchaotic 4y agoWhat’s is the supposed business case for sharing raw, unmodified 42Mpix photos? And if they are still modified later on, and perhaps re-signed, that’s where I would attack.
- TeeMassive 4y agoIf they did things correctly (and I assume a company like Sony did) then each camera will have its own signed certificates attributed to it. Yes you can sign fake pictures, but the signature won't be from the same camera, which is a pretty good mitigation against this.
- karamanolev 4y agoI would assume they can sign both RAWs and JPEGs. I can imagine a hardened coprocessor that can sign things coming from the sensor and image processor, so you get signed RAW and/or JPEG and you can't extract the private keys. Any modification later on means it's no longer signed. Unless I'm missing something, this is pretty good until someone extracts or leaks the private keys. Maybe they came up with something much smarter :)
- CharlesW 4y ago> What’s is the supposed business case for sharing raw, unmodified 42Mpix photos? TFA (TFPR?) answers this. "This technology is particularly applicable for passports and ID verification but goes further in tackling image manipulation in the media, medical and law enforcement fields. For the insurance and construction sectors, this technology will offer a secure foundation for inspection and recording of damage." > And if they are still modified later on, and perhaps re-signed, that’s where I would attack. Even so, the existence of an original capture makes post-capture attacks useless. An effective attack will modify the image before it's signed.
- gjvnq 4y agoPerhaps tools for criminal investigators.
- Schroedingersat 4y agoThe use cases is 'proving' that the picture provided by anyone with the private key is real.
- kelnos 4y agoPresumably this relies on some sort of private signing key being stored securely on the camera somewhere. Do we really trust Sony's abilities to make it impossible to read out that key?
- ChrisLomont 4y agoHave all iphones been broken by state actors? It sure seems not given the resources they pay yo try and crack cases. Why would sony cameras not leverage knowledge gained from iphone security features?
- jitl 4y agoSecure Enclave-like technology is becoming more widely available. ARM can build TrustZone into any general purpose CPU; they’re probably using something like that rather than building their own security module.
- dudeinjapan 4y ago> This technology is particularly applicable for passports and ID verification What if you take a picture of a fake passport?
- karamanolev 4y agoThe goal is to prove that you took an image of something that existed in the state it was photographed in. If you take a picture of a fake passport, you are "proving" this passport existed and looked like it did on the picture. So ... mission accomplished? If they do it competently (or if it can be done) is a separate matter.
- rafale 4y agoUnless you take a picture of a picture.
- karamanolev 4y agoThen you're proving the picture you just photographed existed and looked like it did. Where are you going with this? The utility of knowing a photograph was captured as-is should be know, as should be the ways to physically manipulate things so you capture what you want. IMO the question here should be if 1) this is the best solution for those use cases and 2) if this is implemented in a technically competent way.
- rafale 4y agoI meant the end viewer will not notice it was a picture of a picture of a mountain but will rather think it's a picture of a mountain. Perhaps thinking it's an original work. Under perfect conditions, it's very hard to tell the difference
- Aaargh20318 4y agoI think the use case would be to take a photograph of a person to be used on a passport, not to take a photograph of a passport. Optical validation of a passport from a photograph is impossible, and unnecessary. If you need to remotely validate a passport there are already secure ways of doing that using the NFC chip embedded in most passports.
- Animats 4y agoDoes each camera have its own signing key? If not, "signed by some camera" isn't that interesting. Is there a public certificate registry?
- karamanolev 4y agoGoogle-backed Android devices have a certificate chain where each device has its own certificate, but there's also a root of trust. So you can sign things identifying the specific device, but you can also just verify it was signed by a Google-backed Android device. They might have done something similar. Otherwise, a Sony-wide secret key is still interesting, in some sense, if done competently (which I doubt).
- hengheng 4y agoWould be nice to see this on their xperia smartphones. They've been copying the Alpha look and feel for a while, and I've been wanting to see signed photos on phones for a while.
- Retr0id 4y agoI've been looking forward to it too, extracting the signing keys sounds like a lot of fun.
- zxcvbn4038 4y agoMarketing hype, if I have physical access to the camera then I can get the key and sign anything. It might require specialized tools or training but it is nowhere near impossible. If I am a sovereign entity I can just compel someone to give me the key. I wouldn’t make any life changing decisions about anyone based solely on the presence of a signature.
- TeeMassive 4y agoIt will be very hard to extract those keys while hiding evidence of tempering. Yes that leaves the threat of state actors, but they will probably have they keys anyway and still the technology will be good 99.99% of the time which means good enough for most intent and purposes.
- josephcsible 4y ago> It will be very hard to extract those keys while hiding evidence of tempering. Why do you need to hide evidence of tampering? You're not supposed to need the camera to later verify its signatures. > good 99.99% of the time That's a bigger disaster than not having this technology at all, because now the 0.01% of forgers will be able to say "look, my photo can't be a forgery, it's digitally signed!" and convince a lot of people who would otherwise disbelieve it.
- bradstewart 4y ago> Why do you need to hide evidence of tampering? If I know my camera was tampered with, I can invalidate the signing key.
- josephcsible 4y agoThe forgers aren't going to use your camera. They're going to use their own.
- TeeMassive 4y ago
- russellbeattie 4y agoSony's marketing department and execs should know better than over promise like this. The camera can and will be hacked, guaranteed. Like PlayStation security, it may take a while, but it'll be done. And that's only if they haven't overlooked any obvious flaws. We all know if you have physical access to the hardware, it'll happen. Anyways, ignore the hyperbole. It's just a press release by inexperienced marketers.
- ChrisLomont 4y agoAre iphones cracked? There's a ton of unclaimed money available to anyone that can open them. Ever look at the list of uncracked videogames and how long the top ones have lasted?
- russellbeattie 4y agoYou mean like this? It was literally the first result. https://www.wired.com/story/ios-exploit-jailbreak-iphone-ipad/ https://www.wired.com/story/ios-exploit-jailbreak-iphone-ipa...
- ChrisLomont 4y agoDid you read the article? The article says there might be a flaw, and it would be a huge deal. It does not state those phones are broken. It does not extract keys, which would be needed in the case of Sony cameras to sign fake pictures. Here's a later followup article from Elcomsoft, one of the best security firms around [1]. The key quote: "checkm8 does not affect the Secure Enclave." So no, they did not extract keys needed that would be needed to sign new phots for the Sony phone. And later iPhones have removed this as a vector. It's unlikely Sony made the same mistakes quite old iPhones did, since they have the ability to learn from past security issues. [1] https://blog.elcomsoft.com/2021/05/checkm8-based-extraction-of-iphone-5s-6-6s-and-se/ https://blog.elcomsoft.com/2021/05/checkm8-based-extraction-...
- nixgeek 4y ago
- CoastalCoder 4y agoIt seems reasonable for Sony to stake a bounty on this.
- wilg 4y agoEveryone's skeptical in this thread (including me), but it does seem like a useful thing to do if you can make it practical. Maybe incorporating some of the design elements of a "Secure Enclave" like Touch/Face ID where physical access destroys the private key? Not an expert in this field but curious what can actually be done.
- noodlesUK 4y agoSo this is basically DRM, but straight out of the camera. I can only think of a few situations where attestation that early in the chain would actually be helpful (forensics, biometrics). However, even if they are using a per-camera secret, tamper proof hardware is often not very tamper proof, especially in consumer devices. Even things like SGX have glitch attacks and similar. I question if it would really stop a determined attacker from extracting the key and using it to sign edited images. It’s next to impossible to prevent jailbreaks. Further, a low tech proof of concept solution could be to telecine an edited image onto the sensor. You’d then be able to get a signed image of almost anything without even jail breaking the device
- makeitdouble 4y agoCan’t an insurance company for instance seal the camera they provide to their agents and have a decent confidence that the signed photo are genuine ? The default hardware might be vulnerable to ultra determined attackers but that can probably be mitigated where it’s worth doing so. Basically if the software is really working, solving for the hardware part can be done at the owner level.
- lazide 4y agoThat’s exactly what this is useful for - stopping simple manipulations and forgeries which are a problem now. People have been using digital photos with nothing already for a long time in court.
- O__________O 4y agoWould not surprise me if 10-years from now mainstream consumer devices and platforms attempt to make it impossible to produce content that’s not able to authenticate the provenance of media. Adobe, Arm, Intel, Microsoft, Twitter, BBC, and others appear to be working to make this a reality via the Coalition for Content Provenance and Authenticity (C2PA): https://c2pa.org/ https://c2pa.org/
- enriquto 4y agoHow would it work in practice? How could they forbid me to edit an "authenticated" image (which is still just an array of numbers) on gimp, and publish the modified image as raw data? Wouldn't I be then "producing content" whose provenance is not authenticated? As long as I have a functional C compiler that can read the files on my computer, I don't see how this C2PA thing can be effective at all.
- Schroedingersat 4y agoIf the OS refuses to provide the DRM'd byte stream to non approved software and the hardware refuses to boot an OS you own, what are you going to do?
- jitl 4y agoYou can publish such an image, but you won’t be able to attest “this image came from my camera and is unedited”. It’ll be unsigned, so marked as “possibly edited” the way HTTP websites are marked by modern browsers as “possibly insecure”. The whole point is provenance chain - not to prevent any editing whatsoever or establish “DRM” that forbids the use of all images from a camera or something.
- EvanAnderson 4y agoNikon had this years ago. The signing key was extracted: https://blog.elcomsoft.com/2011/04/nikon-image-authentication-system-compromised/ https://blog.elcomsoft.com/2011/04/nikon-image-authenticatio...
- bradstewart 4y agoThe ability to have unique per-camera keys with tamper-resistant storage (that can "self destruct" when it detects voltage glitching, etc) has come a long way in the last 11 years.
- SXX 4y agoI hope this time Sony's getRandomNumber function will return something other than 4: https://youtu.be/LP1t_pzxKyE?t=248 https://youtu.be/LP1t_pzxKyE?t=248
- rekabis 4y agoLooks like Polaroid is still king when it comes to unforgeable photographic evidence.