4 ms·

This should be a top comment - this kind of tracking is not possible at scale and probably won't be for a long time. Identifying a person in a group of 100s/100

by rm999 11y ago

This should be a top comment - this kind of tracking is not possible at scale and probably won't be for a long time. Identifying a person in a group of 100s/1000s has been a solved problem for decades. Consistently identifying a single person from a database of 100Ms/billions of people who are constantly gaining weight, growing facial hair, etc is impossible with today's technology; false positives will become an exponentially bigger problem as you increase the candidate space. Perhaps one day increased camera resolutions will allow trackers to identify something humans can't, like retinal scans. This would allow cameras to basically "fingerprint" people, and will be super creepy.

IanCal 11y ago
I think that actually as you increase surveillance the problem gets simpler. If you're always in view of a camera then I can simply track your motion and I only need to recognise you once. Given that I've seen you in your car, I can already narrow things down significantly. > Consistently identifying a single person from a database of 100Ms/billions of people I don't need to, there aren't billions of people who are likely to be at the closest ATM to your house. > constantly gaining weight, growing facial hair, etc But I see you several times every single day. > false positives will become an exponentially bigger problem as you increase the candidate space. To be honest, I think that this is the most likely thing to scare people, that you won't be allowed to fly because a computer made a mistake and everyone trusts it or can't override it. That the police get called when you go to pick up your son because you're the same height and rough build as a child molester who was released from prison yesterday
nitrogen 11y ago
There are already devices that can rapidly scan irises of a crowd. It's also possible, as the sibling comment suggested, to correlate multiple datasets (wifi probes, unique phone CDMA/GSM/LTE IDs, location, behavior, physical limitations like maximum velocity, etc.).
finance-geek 11y ago
"people who are constantly gaining weight, growing facial hair, etc is impossible with today's technology;" I think we can assume many agencies have access to ALL Facebook data, Photostream data, etc. With that it is very much possible, dare one say, easy?

by rm999 11y ago

IanCal 11y ago

I think that actually as you increase surveillance the problem gets simpler. If you're always in view of a camera then I can simply track your motion and I only need to recognise you once. Given that I've seen you in your car, I can already narrow things down significantly. > Consistently identifying a single person from a database of 100Ms/billions of people I don't need to, there aren't billions of people who are likely to be at the closest ATM to your house. > constantly gaining weight, growing facial hair, etc But I see you several times every single day. > false positives will become an exponentially bigger problem as you increase the candidate space. To be honest, I think that this is the most likely thing to scare people, that you won't be allowed to fly because a computer made a mistake and everyone trusts it or can't override it. That the police get called when you go to pick up your son because you're the same height and rough build as a child molester who was released from prison yesterday

nitrogen 11y ago

There are already devices that can rapidly scan irises of a crowd. It's also possible, as the sibling comment suggested, to correlate multiple datasets (wifi probes, unique phone CDMA/GSM/LTE IDs, location, behavior, physical limitations like maximum velocity, etc.).

finance-geek 11y ago

"people who are constantly gaining weight, growing facial hair, etc is impossible with today's technology;" I think we can assume many agencies have access to ALL Facebook data, Photostream data, etc. With that it is very much possible, dare one say, easy?