Add random noise does not make noise analysis pointless? #4
Loading…
x
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
As evaluating fixed patterns is difficult or even some biased random patterns are impossible to correct?
Like manufacturer/OS could implement it for privacy but at the same time they want the best quality?
In general are there physical and software methods to avoid PRNU?
Or only if a few pixels are concerned, then could just interpolate them, as I did in another context:
In theory could just mean added noise and it would converge to 0, no?
Related to Benjamin-Loison/android_packages_apps_Aperture/issues/7#issuecomment-2091275633.
See section V. of src/commit/ec9f6e7acaaeb6b1ffe14d3841e87bf8cc7e526b/articles/Digital%20Camera%20Identification%20from%20Sensor%20Pattern%20Noise.
In theory the manufacturer could establish the PRNU of each camera to then recognize what device thanks to a given image and find who buy it etc.