SoylentNews is people
SoylentNews
HMD's Nokia 2.3 has been announced for sale in the US. This low-end phone is just $129 but still manages to look like a respectable device.
[...] it runs stock Android with no crapware.
[...] There are some nice extras here, too, like a micro SD slot, a headphone jack, FM Radio support, and a dedicated Google Assistant button on the side.
[...] There are some downsides that come with the low price. First, there's no fingerprint reader. The only biometrics are a selfie-cam-powered face unlock feature, which can't be that secure since it only does a 2D face scan. Second, you're getting the old micro USB port for charging, instead of the newer, reversible USB-C, which is a shame. This also means there's no quick charging, and instead you get a pokey 5V/1A charger. Third, there's no NFC, so you won't be able to tap-and-pay at the register. Fourth, it only comes with Android 9 Pie, although an upgrade to Android 10 is planned at some point.
https://arstechnica.com/gadgets/2020/01/dirt-cheap-nokia-2-3-comes-to-the-us-for-just-129/
Only downside [for me] is the fact that it actually has a selfie-cam-powered face unlock feature. How about you?
(Score: 1, Insightful) by Anonymous Coward on Thursday January 23 2020, @06:33AM (2 children)
"which can't be that secure since it only does a 2D face scan"
The author doesn't understand imaging nor math. Every consumer phone sensor is a planar array. Sometimes multiple sensors are used. Always, any 3d data is retrieved by correlating pixels between frames or between sensors or both. Generating a 3d pointcloud from a 2d video stream is how they all work.
Further, assuming that 2D is inherently less secure is ridiculous. That's like saying block encryption is more secure than bitwise stream encryption. The devil is in the implementation.
(Score: 2) by kazzie on Thursday January 23 2020, @09:23AM
It could be 3D (or stereoscopic) if there were two front-facing lenses, but in my limited experience you only find multiple lenses on the back.
(Score: 3, Insightful) by Immerman on Thursday January 23 2020, @02:29PM
Generally speaking, single camera facial recognition can be easily fooled by a photograph taken by another camera - a photograph of a photograph looks just like a photograph of the original subject. Stereo+ cameras in contrast can build a 3D model of your face, and thus require a 3D model of your head to fool, which is much more difficult for most people to acquire.
Of course a lot comes down to implementation, and you could potentially even require users to rotate their head to build a 3D model with one camera, but that could still be fooled by a video of you doing the same.