You can login if you already have an account or register by clicking the button below.
Registering is free and all you need is a username and password. We never ask you for your e-mail.
I don't think they have studied the gender gap properly. Makeup acts similarly to the IR lighting that they talked about. It changes the bright and dark areas of contrast and that can fool the technology.
The fact that they just showed the gender and race data with no explanation other than to imply that facial recognition is sexist and racist was a bit disingenuous to me. Had they taken any time to actually examine the causes - which she diligently spent time on detailing how the dots and meshes were created (same principle) - the data would make sense.
I hate all of the implied "technology is racist" tropes.
I wonder if anyone has ever looked at camo paint as a method of confusion.
Yeah. But I'd heard all the good points already. If I had anything better to do (at work all day) I'd be pissed they crammed so much garbage into a lecture and wasted my et minutes or whatever.
view the rest of the comments →
[–] anamazonslittle 0 points 1 point 1 point (+1|-0) ago (edited ago)
If you want to skip all the bullshit "gender is a spectrum!" and "facial recognition discriminates against black people!" go to 19:00
1) wear a mask. Some facial recognition software can be tricked with a printout of a face.
2) wear sunglasses. It might still detect your face but will have a much harder time determining who you are
3) make silly faces
4) a paparazzi scarf/visor/reflective glasses
5) IR lights under the brim of your hat that shineonto your face
6) a lot of really bad makeup (like you're going to a convention kinda ugly)
7) get privacy laws changed.
[–] AmaleksHairyAss [S] ago
There was one useful lesson in the gender bullshit: You may be able to use makeup to look different from the probable training set.
[–] tidy_athiest 0 points 1 point 1 point (+1|-0) ago
I don't think they have studied the gender gap properly. Makeup acts similarly to the IR lighting that they talked about. It changes the bright and dark areas of contrast and that can fool the technology.
The fact that they just showed the gender and race data with no explanation other than to imply that facial recognition is sexist and racist was a bit disingenuous to me. Had they taken any time to actually examine the causes - which she diligently spent time on detailing how the dots and meshes were created (same principle) - the data would make sense.
I hate all of the implied "technology is racist" tropes.
I wonder if anyone has ever looked at camo paint as a method of confusion.
[–] anamazonslittle 0 points 1 point 1 point (+1|-0) ago
Yeah. But I'd heard all the good points already. If I had anything better to do (at work all day) I'd be pissed they crammed so much garbage into a lecture and wasted my et minutes or whatever.