4 ms·
Am I just a crazy cynic or are ZK proofs here just a buzzword. Like, how is this any different than having each camera equipped with a vendor controlled key an
by nixpulvis 1y ago
Am I just a crazy cynic or are ZK proofs here just a buzzword.
Like, how is this any different than having each camera equipped with a vendor controlled key and then having it sign every photo?
If you can spoof the sensor enough to reuse the key, couldn't you spoof the sensor enough to fool a verifier into believing your false proof?
- injidup 1y agoYou take a photo of an AI generated photo. What's yr proof worth then?
- a3w 1y agoYes, IIRC if you measure an image signal, i.e. here: image, that is at twice the resolution of the sensor you use, there won't be any artifacts. Nyquist–Shannon sampling theorem. But, if the sony sensor also measures depth information, this attack vector will fall flat. Pun intended.
- Barbing 1y agoThe technique you describe has seen some use by some major vendors for 5 to 10 years right?
- nixpulvis 1y agoNot sure.