Windows Hello's facial recognition might not be a good thing with twins? [It works against twins]
Apologies to Mr. Bean, but someone really does look like him out there.
metro.co.uk/2015/09/17/you-wont-believe-how-much-this-teenager-looks-like-mr-bean-5395192/
What are the chances of facial recognition unlocking the phone/device because of a twin or doppelganger?
Comments
-
Hi @wkleem,
Just because twins or doppelgängers are identical does not mean they have the same facial structure all the way through, the face scanners can tell them apart on much deeper levels that our brain/eyes cannot work on. Our eyes are actually not the best at detecting details but rather specialized at patterns instead to connect to stored images in our brain. That's why when certain people lose weight, not everyone can tell right away that they did, your brain is processing on stored images, not what it sees right away. In fact, there are also blind spots that we naturally have that doesn't see anything and not to mention, magic experiments have proven that we don't pay attention to the details.
One news site did test it with several sets of twins and Hello worked correctly, here's the results: http://www.theaustralian.com.au/business/in-depth/windows-hello-can-identical-twins-fool-microsoft-and-intel/story-fnw66tov-1227490164701
0 -
Fascinating! Thanks for the link.
0 -
You're welcome!
0 -
Hi, I wonder what the margin for error is. For example, there are or were some twin that were so alike to be virtually indistinguisable. I was reading about the Toni twins.
http://www.people.com/people/archive/article/0,,20065619,00.html
September 1975
"Lila and Ella have discovered that they are truly two peas in a pod. "A member of the test team told us that our bone structure and fingerprints are so similar that we could be one person. Our blood pressure, cardiogram, cholesterol level—everything is the same." The tests have had one unexpected side effect: the two women have begun dressing alike again and wear their hair in identical frizzes."
0 -
Hi @wkleem,
That statement was made back in 1975 where testing were really primitive compared to today. I'm willing to bet if the said testing was done with today's technology, it won't output the same answer.
As for margin for error, it is probably very low but only Microsoft and Intel can answer that question for you.
Check this video out where they show what your face look like under a different light spectrum:
https://www.youtube.com/watch?v=o9BqrSAHbTc0 -
Hi
I saw this video some time ago. It never would have occured to me that two identical looking people, indistinguishable to the naked eye, would have enough differences for the camera to know the difference between them!
Thanks.
0 -
It's amazing, isn't it?
0 -
It's difficult for us silly humans. We rely on their senses too much, and worse, steadfastly believe they are effective tools for perceiving or learning of the actual physical Universe, that they tell us the truth. "I Know What I See!" Nothing could be further from the truth. Our senses are grossly inadequate, just sufficient enough to allow us to survive and procreate.
0 -
Kind of like a Dunning-Kruger effect for perceiving the world around us, eh, @MrC?
0 -
Indeed, and more. Its hard for us to fathom, for example, that our visual spectrum is so astoundingly narrow relative to the full electromagnetic radiation spectrum. Even when explained reasonably well, folks have a hard time accepting that most of the phenomenon we think of as "light" is not visible to us. Or that some creature like a dog can differentiate and detect scents a million times greater than we can. Or that a bird "see" magnetism lines circling the earth. Or that a shark can taste blood from miles away.
We're so pathetic. :-)
0
