In article <4i4aI.269490$ic4.2...@fx47.iad>, JF Mezei
<
jfmezei...@vaxination.ca> wrote:
> > you are confusing a depth map with face id data.
> >
> > apps can get a depth map.
> >
> > apps *cannot* get facial recognition data.
>
>
> With TouchID, apps, even kernel level code cannot process TouchID data
> because it is encrypted by the button itself with the decryption kety
> known only to secure enclave.
>
> With FaceID, the sensor data is usable by applications via APIs. (to
> enable useful features such as animating the poop emoji to follow face
> movements.
like i said, you're confusing a depth map to track facial expressions
with facial recognition data that can unlock the device.
apps can get a depth map, which poses *no* security risk.
apps *cannot* get facial recognition data, nor can it be spoofed.
> With TouchID, there is a hardware firewall since the button itself has
> the chip to encrypt the data stream that flows from ribbon cable to
> motherboard to CPU and to secure enclave.
true, and they must be cryptographically paired to guarantee a secure
path.
> With FaceID, the protections are more software based because the data
> from sensors needs to be available to the CPU to animate the poop emoji.
false.
> This starts to matter when you're the NSA trying to unlock a phone and
> using tools to inject a face map of the owner of phone you want to open.
it doesn't matter since it's impossible by design, for both a face or
fingerprint.