It’s my understanding they don’t work against all kinds of depth scanners and that there are some researchers/agencies who are using gait tracking which is promising and doesn’t require facial recognition. I assume it might require mass surveillance and that gaits may be localized, but the general sense I have, is we shouldn’t get complacent and every time we think we are safe, we are probably not being creative enough.
I'd like to see if that actually works as an adversarial technique to gait tracking, I would assume it wouldn't - given that a gait has a lot of variables, a pebble in the shoe should just tweak one or two, and only reduce significance, but not throw a wrench in the identification entirely.
Part of me wants to say "duh", but studies that show obvious results are still useful.
I had a similar thought when I saw the paper last week that showed masks that don't cover the nose are less effective than those that do. On the one hand, why did we need a study to show this. On the other hand, now we have a study confirming it.
We are just not there yet. For all the people that I can't recognize initially, I do recognize many. And if I can do it, some supercomputer probably can in the future, and the future is coming fast.
Good. Facial recognition coupled with mass CCTV is going to force us to make _hard_ choices in respect to privacy, and I don't think any society on earth is truly prepared to deal with that yet.
Again, as always, what you call Britain the reality is that its centers around a small number of cities and primarily the heart of London. And again, a large percentage is privately owned and not wired into the kind of Person of Interest system that people appear to have conjured up in their minds.
Similar views can be found between Sydney and the rest of New South Wales, and also between Auckland and the entirety of New Zealand (the kiwis affectionately refer to Aucklanders as JAFAs - Just Another F$%king Aucklander, because half the country lives there).
One would suspect such complaints about the friction between centralism and federalism are widespread across the world.
I always chose to believe it was Just Another Friendly Aucklander. I knew I was lying to myself, but it makes you feel slightly better than acknowledging 2/3 of your fellow citizens don't like you.
I doubt it is personal. I live outside of my state's major city (Sydney) and feel no personal dislike to anyone in Sydney. It is just difficult to understand why someone would want to live there. And I think people outside major urban centres often feel ignored by politicians and this leads to resentment.
Brexit was caused by London to some degree. Their financial industry creates a huge demand for the pound, which hurts exports for the rest of the country. This has dampened standards of living outside the Capitol for a long time now. This phenomenon is also called dutch disease. https://en.m.wikipedia.org/wiki/Dutch_disease
Could also be rephrased as "Do you mean that the all the CCTV surveillance in the City is decentralised and only accessible by warrant to each system owner?"
In the City, I can easily imagine there's still a single entity behind it, just one that's not accountable to the public. No idea whether that's the case, though.
Anyone running CCTV needs to register with the Information Commissioner and there are some limits to what can be done with the footage or how long it can be kept for.
What do you mean by that? Prevention, detection or prosecution? Genuinely interested. For certain crimes (I'm thinking of the Salisbury poisonings for example) CCTV seems, on the face of things, pretty central to finding out what actually happend.
This is a classic example of twisting statistics to fit a narrative. That match had ~65k attendance with ~2k false positives for sake of argument. thats < 3%. Not great, but then consider the system is going to scan a vast portion of those in attendance which a human eye alone could not do. Even with the false positive rate, in terms of making human jobs easier, its surely a benefit to scan 2000 false positives than some large portion of 65,000 when checking for unsavoury characters.
From the article: "But of these 92% – a total if 2,297 – were incorrect, with just 173 providing ‘true positive alerts’."
Those are frightening odds when you consider the impact such system could have on those "false positives", eg. possible detainment until your identity is confirmed.
The police may have had the resources to check each case in person during the trial but you can imagine what could happen once there is a push for more wide adoption of these systems as a way to save on police resources.
I would also love to see the proportion of people with dark skin tones who were flagged as false positives.
Then what happens to the pictures of non-confirmed positives and negatives? Are they purged from the system automatically or are they kept around for a while? What recourse have the false-positive to not be harassed in the future?
Let's face it, this genie is not meant to stay in the bottle and it's a policing dream to be able to automatically identify potential undesirables, wherever they are. Why not at the entrance of banks and other sensitive places? Like sports clubs and parks and swimming pools. At every corner of the streets maybe?
> Even with the false positive rate, in terms of making human jobs easier, its surely a benefit to scan 2000 false positives than some large portion of 65,000 when checking for unsavoury characters.
who says it sensible to check them at all? Basic probability. Let's say the false positive rate is 1%, and there's a criminal among every thousand. That means on chance alone you'll get ten false positives for each actual hit. (the low base rate makes the false positives much worse than they sound, same reason you shouldn't panic if you get a positive test for a rare disease).
That is a ton of work for nothing, a lot of people being stopped without reason, when the saner alternative may have been to not pester anyone at all. Discovering criminality comes at a cost to people who are falsely dragged into it, and creates opportunity cost for the police.
Do you base this based on numbers or gut feeling? I've always felt unnaturally safe in London despite it being a massive sprawl of people-filled chaos.
You should certainly feel safe in London, but you can also feel safe in Paris, Berlin or Amsterdam. That was the argument, your re-framing quite deceptive.
Numbers are difficult to compare since data collection mechanisms are subject to change, but here you go:
A few years back I read a statistic in the economist that said CCTV coverage in areas reduces crimes at night to the same degree as better lighting. Sure, there is an effect but it can easily be achieved with other methods as well.
This always requires the question: less safe for who?
Per that article, in 2010 the vast majority of violent crime offenders were black - and I wouldn't be surprised if the victim profile would look similar.
Pure anecdata, but this was 90% of the news I saw. Gang stabbing in park. Gang stabbing at party. Gang shooting at club.
If you and I were to travel back in time to the high trust society before mass surveilance & enquired our police wouldn't deny they held the the monopoly on violence. My friends in the forces of today tell me they feel under siege most days.
Apologies, "London is the knife crime capital of the world." is part of the venacular here akin to "The man who broke the Bank of England".
With _the_federal_government_ ensuring that 5G covers every inch of the territory enabling economies of scale for low cost SOCs and pricing tiers, the days of WiFi for any kind of mass deployment (doorbells, ATMs, street cams, etc) are ordinally numbered backwards from a small integer. With all fixed location cameras connected in that way, he who controls the 5G controls the world...
I watched two DW documentaries on coronavirus in China[0][1] -- [1] is now private but you can see what the page looked like on the wayback machine[2]. What I remember from one of the videos (not sure which) was that they had facial recognition with masks working, they discussed it and there was some b-roll of facial recognition cameras set up inside a mall(?) and other places.
The other commenters who have noted that there are lots more variables involved (height, gait, etc) are also right of course.
I think China is a world leader in this field of technology, and I fully expect them to have solved this, for better or (almost certainly) for worse. Imagine if you had an organization with virtually unlimited local currency willing to pay researchers/smarter-than-average/average programmers to solve the who-is-that-in-this-picture problem, but also you had no regulation to answer to with regards to personal privacy and/or data access ("it's for national security" gets you close to no oversight in western nations, but not quite China-level). Consider the problem mostly solved. Even if it wasn't, the fact that just about every citizen carries a cell tower/GPS beacon with them should be enough to help narrow down the search space dramatically, even if your algorithm/methods were dogshit (it isn't).
China has created a real life 1984, and tech helped them get here. I don't know whose (including me) conscience that's supposed to be on but just putting it out there.
Part of the 1984 method is that even when techniques don't work, they lie and say the techniques do work, to scare the populace. See also the polygraph machine
Any information regarding that which came from China just gonna be biases. They obviously gonna try to sell their mass surveillance systems as greater good that eliminated crime completely.
In real life totalitarian country don't need ro worry about unsolved crime since there ia no one to report on it. Anyone talking about any kind of crime on social media just gonna be slienced or just censored to begin with for spreading rumors.
I remember reading that the Chinese commonly used far more advanced algorithms that looked at not only face but height, posture and walk to identify even without the face, or with a partially obscured face.
Basically like browser fingerprinting. Use a lot of indicators that are insignificant on their own to create one powerful method of identification.
The problem with those is that it's really hard to get around them. Changing one of the factors isn't going to help. You'd have to identify the most significant ones to change those (if that's possible).
It sure breaks mine. I saw my cousin in a mask in a restaurant, she recognized me and I didn't know her until she spoke.
Maybe once we don't look at a person and think "Facemask!", once it becomes normal, we'll retrain to identify folks by eyes, head shape, body language etc. And quit relying on nose and mouth so much.
There is a bigger impact on the biometrics industry. Face recognition (because of the masks) is the most obvious one, but there are also fingerprint scanning and voice recognition which took a hit. As a viable alternative, typing biometrics could be the safe & user-friendly way to go (recognising people by the way they type). You can see more on this topic on the TypingDNA blog: https://blog.typingdna.com/covid-19-pandemic-challenges-the-...
That is not the case. I played around with this technology for a bit and it stores more than just one pattern in order to better understand how you type.
Fingerprint scanners are legacy tech. The cool kids now are using palm vein scanners - great CER and non-contact, you just hover your hand an inch above.
The problem is convenience. I have to press the elevator button every day. After I disinfect my hand for the 5th time that day I use the stairs. Same with the fingerprint scanner.
If I recall correctly several medical conditions can be detected by retinal scans, which has historically made widespread use of the technology undesirable.
EDIT: Sorry, it seems that it's more that medical conditions make it unreliable
> Certain eye-related medical conditions and diseases, such as cataracts and glaucoma, can render a person unable to use retina-scan technology, as the blood vessels can be obscured.
Haha! Wow, the part where there is a device down the pipe that acts as a mini laboratory, analyzes the fecal matter passing through, and sends you medical alerts on your smartphone… I could totally see this exist in the future. Identification could probably be done by sequencing the gut microbiome to avoid the camera taking pictures of your ass.
Out of interest has anyone tried setting up an alternative appearance in Apple’s Face ID with a mask on yet? I only have my normal appearance set up and a mask is enough to stop that working - which is pretty unsurprising. I didn’t realise you could have an alternative appearance set up, curious how well that’ll work - can it accurately learn and then recognise with a mask on?
The question that comes to my mind is, how much worse did you make your security by enrolling with a mask. Because if it recognizes you with a mask 50% of the time, how many non-yous with a mask would it recognize as you as well?
I wouldn't touch that link with a 10 foot CAT6A cable.
Considering that license plates exist specifically for the purpose of recognition (originally by human observers), I don't think it's news in 2020 for anyone to be identifying them via machine vision. If you have a Raspberry Pi and a Pi Camera, you too can participate in this kind of surveillance.
"Antifa" seems to be a symptom of fascism and racism -- like an elevated white blood cell count in response to an infection. I'm not sure the organization even exists. It certainly wasn't a major part of the national dialogue prior to that fatal rally in Charlottesville.
Antifa did exist in the 1930s as the militant wing of the German Communist Party [1] and whilst working in London in 2012 I saw many organised protests of theirs for different causes than in the prognosis.
If the organisation existed today given the majority have been mobilised through steadfast polarisations, a multitude of grievances and not as cover for potential infiltrators, why would they admit it? It is much more convenient to remain nebulous.
The link adds credence that this is more an insurrection rather than "proud mothers" who it seems were organised to act as shieldbearers to protect the bombers of Americas federal buildings [2] and not their children to whom they hold a duty. You can see this mass organisation by self-identified antifa-associated accounts #PortlandMoms and scroll down.
The parent site itself, is that one of the splinters of the forum that reddit banned?
I can and have read Wikipedia plus many forums on the extreme left and right's opinions Antifa and if it exists. But quite frankly in this conversation it is the equivalent of X that does not intersect with the government.
And I know about OpenALPR.
And I know facial and license recognition has been around since the 60s.
It's implicit we are talking about the data collection and large scale database and organisational control of that data.
I don't know why that link scares you, but you should learn Tor or AdBlock or Tails or VM's, I don't mind paranoia but lazy paranoia is, well, lazy.
It's time to switch to gait recognition I suppose. Same cameras, different algorithms. Where cameras can be upgraded they can be upgraded to also capture some IR bands, in which you can see through (some) masks, clothing, and even some plastics. Algorithms could stay roughly the same in that case.
