Facewatch 'thief recognition' CCTV on trial in UK stores | Hacker News Reader

Facewatch 'thief recognition' CCTV on trial in UK stores

bbc.co.uk

51 pointsSimplyUseless78 comments

> We later discover my thick-rimmed glasses are the problem. It's something easily fixed by adjusting the system settings, Simon tells me.

Facial recognition is a hard problem; allowing for thick-rimmed glasses is not 'easily' fixed.

Good, corporations have been allowed to go too far with their surveillance.

Well, you can't admit that to a reporter/the public...

So, how would a reformed thief go in to buy basic goods once this becomes more widespread?

Do we no longer believe that someone can pay off their debt to society and become a useful member of said society?

Nope. Fuck them anyway, and fuck any twins of said thieves. Fuck them doppelgangers, and any person resembling them. And any false positives, for that matter.

I, for one, welcome our precog overlords.

So, no, you don't believe in someone being able to reform. The rest was not really necessary.

Let's say that 90% people of thieves successfully reform and become honest people.

Assuming that 1% or less of general population are thieves, it still would mean that a previously known thief has a 10% chance of being 'unreformed', and that they are a 10 times bigger risk than the average person; and it's not prohibited to use that information to safeguard yourself.

We're talking here about thieves that hadn't been caught by the Police.

Maybe the way to handle it could be that if you're a thief and had a change of heart, report to the local Police station and agree to serve your time in exchange for getting your face removed from the database.

Wait we are talking about, data that could be easily manipulated because "I don't like him" ?

Is it in the interest of grocery stores to exclude customes without a good reason?

You're missing the largest issue. If the customer is just a dick and did nothing illegal the store owner out of spite might put him or her into the database to be flagged in every other store.

Hopefully the significant fines would keep this at bay. In all honesty kicking the store off the system would be the best way to stop it.

I see your point and you're right, this is a potential problem - mostly because networking this system would make consequences explode out of proportion. Maybe some appeal system could resolve that?

As a side note: I support the right of store owner to punish dick customers. I believe society needs to have stronger mechanisms to penalize douchy behaviour.

> As a side note: I support the right of store owner to punish dick customers. I believe society needs to have stronger mechanisms to penalize douchy behaviour.

I agree with this, but I don't believe that a store ought to be able to farm out this punishment to every other retailer that subscribes to this system.

If a customer is being a dick, bar him from your establishment; don't subscribe them to be discriminated against and harassed everywhere else.

I totally agree with you. Banning a bad customer from your store is one thing, banning them from all other (subscribed) stores as well is way too much.

No, but their idea of a good reason might be "they're middle-eastern", "they're black", "they're old", "they're young".

The database only has convicted thieves in their Database, so in theory they've already served their time for their crime. The problems is that they are then put on an apparently un-auditable list maintained by a private company, and discriminated against.

> The database only has convicted thieves in their Database,

From this article:

> and other potential offenders

And from this article about a different system:

> Police forces are already using face-recognition to identify people, much as they might with DNA or fingerprints.

> A national police database contains up to 18 million "custody suite" photos - hundreds of thousands could be of innocent people.

Many of those people will be those arrested for an offence, but not cautioned or tried. (To be clear here: they really didn't do the crime, not just they weren't convicted of a crime).

Thanks for the clarification. Not a point in favor of the system, though.

The article makes no reference to how the positive result may be used. For all we know it might only be used to allow shop associates to further monitor certain individuals.

Although I'm sure some local corner shop will try and "ban" a positive result and make the headlines. That's inevitable. But for the larger shops, I doubt they will make it known that they know.

From the article:

> "The people who are on the list are not guilty until they've been prosecuted and taken to court, and the system makes that very clear", Simon says - and under the Data Protection Act "if anyone misuses that data there are very significant fines".

Probably someone will follow you around the shop or ask you if they can help you instead of you not being let in into the shop. Those are already pretty common tactics.

Those tactics also don't feel nice though (I experienced them first hand, when I was in a foreign country and was a bit nervous in shops - not because I was shoplifting but because I wasn't sure wether I would be understood by the cashier).

So, they are put on a list and discriminated against _after_ paying for the original crime, on the belief that they are incapable of reforming.

Personally, I'm not comfortable with that concept.

The stats are not in the favor of your argument:

"Property offenders were the most likely to be rearrested, with 82.1 percent of released property offenders arrested for a new crime compared with 76.9 percent of drug offenders, 73.6 percent of public order offenders and 71.3 percent of violent offenders."

http://www.nij.gov/topics/corrections/recidivism/pages/welco...

The stats don't matter, IMHO. It could be 99% of them who re-offend; to segregate and discriminate against them after they have paid for their crime is a strike against the basic human rights of those people.

No it isn't. If 81% of them are re-arrested (which means much more than 81% commit the crime again), a little inconvenience and suspicion is fully justified.

Previous conviction record is not one of so-called protected characteristics (gender/race/religion/etc), it is completely legal to discriminate based on that.

Basic human rights include a lot of things, but the list of those does not include a right to be trusted despite your prior actions, and it does not include a right to be served in particular location - those are privileges or interpersonal relationships that may be granted or withheld as the others believe is best for them.

> it is completely legal to discriminate based on that.

Unless you're breaking the rehabilitation of offenders law.

https://en.wikipedia.org/wiki/Rehabilitation_of_Offenders_Ac...

Yes, most jurisdictions have some kind of expiry for criminal record.

However, for good reasons, this doesn't happen automagically at the end of a prison sentence, it requires much more time than that and many thieves (the profession having a high rate of re-offending) spend most of their out-of-prison lives without this expiry/rehabilitation occurring.

As long as there are adequate safeguards for incorrect data (as should be provided by the DPA in the UK), I don't see any problem with this.

It's no different to companies choosing not to do business with people based on credit history, financial standing, etc.

Imagine, for a moment, that I shoplift as a teenager as a drunken dare. I'm caught, and pay the fines and a bit of community service.

Or, perhaps I'm just suspected of shoplifting because one of my friends did shoplift while I was with them, and I'm added to the list. The doubt is reasonable enough that it wouldn't get me convicted, but it's enough for this system to add me and keep me in it.

For the rest of my life, I would be treated as a criminal by shopowners, even if I never commit another crime.

Any list like that sure sounds like slander/libel. If you are treated as a criminal by a shop-owner because someone else is making claims about your reputation, this is the exact type of situation that slander/libel laws were created for.

Also, I could be badly mistaken, but isn't slander strongly enforced in the UK? Wasn't there a scandal involving forced apologies?

It's down to the retailer if they want you on their property, and that's how it should be.

If they opt for a stricter system (which would cover the edge cases you describe) that loses them some genuine custom, that's their loss.

The target is of course persistent offenders, not the kind you mention.

If a "reformed thief" is in my store, I would love to know about it and provide him a friendly escort. All the way to checkout.

He's of course welcome to never come back again. Seems like a fair trade off. Not my job to feel bad for criminals, it's my job to put food on the table legally.

The history of people who enter your store is none of your business. Or are you really suggesting that you support segregation?

Everything in my building is my business. That's why it's called... a business.

A business does not get a free pass to act however they want in many parts of the first world. The classic examples are discrimination based on religion, disability, or race.

There are a bunch of characteristics based on which we shouldn't discriminate people, and in the first world it is illegal to do so.

Your reputation is not like that, everybody should treat you according to your reputation, wether it's good or bad.

In the UK, we have the Rehabilitation of Offenders Act to give convicted criminals the ability to reform and reenter society.

I wonder if the CV Dazzle[1] style will take off in the UK. Maybe we'll finally get the exact cyberpunk future promised us by Gibson, etc.

[1] https://cvdazzle.com/

No need to go that far: every system I'm aware of fails when there are no eyebrows. (I'm in the FR industry.)

Your eyebrows must be [<-------- this -------->] long to enter this store/view this content.

So Whoopi Goldberg might be some master thief and there is nothing that can be done?!

I feel like shaving off your eyebrows is actually going further than putting some make-up on.

Like Tor and other security/privacy measures, it only really works when it's generally adopted; otherwise people in the style will stand out on their own.

I believe someone working on Tor (or similar privacy tool), said that a privacy system is only good if it's used by people to share photos of cats.

If a tool is only used by shady people for shady means, then the tool itself (and use of it) is suspicious. But if it's used by many average people to do mundane things like sharing photos of cats, then subversives can use it too

There are clubs in most British cities where the CV Dazzle look wouldn't look out of place. (From experience, while travelling to and from the club one certainly looks out of place.)

https://www.google.co.uk/search?q=cyber+goth+hair&tbm=isch

I wonder if places using systems such as this will have to declare it plainly?

If they do, seems like there might be some backlash, with people preferring not to go to businesses starting down the path of using this sort of tech and spending more time in those which could, but choose not to, use it.

Assuming of course it doesn't end up just completely ubiquitous like web ads etc, the market could force tech like this into the high-risk margins, i.e. high-end jewellery shops etc, rather than it eventually popping up in every pub and high-street shop.

I will say though, having written that, I do think it's likelier it'll just end up ubiquitous.

> I wonder if places using systems such as this will have to declare it plainly?

There is a lot of law about data protection in the UK & EU. So, yes.

Some pubs/clubs already require things like fingerprints to be taken before going in, it doesn't seem to stop people going there. If anything, something which is less effort for the customer (like the Facewatch system) would probably be preferred

It won't stop everyone, some people don't care, might be socially pressured to just going in anyway in a group, etc. But for customers who do care, it could actually work in the opposite way too.

If a club wants to take your fingerprints, it's usually after you've already entered the place, having possibly queued, at a point where it's very annoying and socially awkward if in a group to about-turn and say "you know what, no, let's go elsewhere".

If on the other hand there's a sign outside declaring "we use Facewatch here" then it's actually much easier for customers for whom that represents a problem to just continue walking past, and not be drawn in in the first place.

It may be that simply not enough people care about having their biometric data read and stored in databases etc, or that it'll be so ubiquitous that there's no alternative but to go along with it. But like I say, it's possible, if either of these things isn't true, that the market will suppress technologies like these becoming 'just the way things are'.

Again though, I honestly don't think that is likely looking at what's happened with web tracking, CCTV and the like. It's possible this, however, is a bit more 'in your face' (pun intended) and will spur a greater negative reaction in people.

CCTV didn't change much the game, I don't think this will either.

Widespread facial recognition is going to add a whole new slew of data for GCHQ/NSA to data mine. Not to mention advertisers tracking exactly where you walk in shopping malls and shops.

From my understanding, UK law requires that you be informed if subject to facial recognition, but, analogous to online tracking/advertising, does not require your permission to begin it.

Sigh, gets out my rimmed glasses

Would informing you be different from the 100s of "CCTV in operation" signs you walk past every day?

The requirement to inform is easily met with a nice sign posted at every doorway saying, "smile, you're on camera (facial recognition)". It will easily be accepted by the masses who simply want to get in, buy their stuff, and get out.

> Widespread facial recognition is going to add a whole new slew of data for GCHQ/NSA to data mine.

Less than you think.

We're all walking around with beacons in our pockets already. They already have the ability to "follow" a person to within a square mile, and that individual is far more knowable (i.e. less false positives) than facial recognition.

Between IMEI spotting and license plate readers surrounding central London, they have a real time SimCity-esk map of the city. The facial recognition stuff might be the cherry on top, but it is hardly going to be a major gain.

PS - And as far as I know they already have facial recognition cameras as most major tube stations.

What happens to a face after it's in a database, or multiple databases, gets arrested, spends time in jail, and then goes straight? Will it be as impossible to get off these databases as it is to get off the no fly list?

And that has nothing to do with FR and its technology. The contents of an FR database and getting items removed is purely human politics.

This technology, deployed, is without a doubt a political issue. Divorcing what we work on from how it's used is irresponsible. And I'm not saying don't develop technology, but we have to recognize and consider how it fits into society. We have to discuss.

The Stasi recruiting human spies from among the general population was both a logistical question and a political issue. The NSA weakening encryption, or parking on backbones or datacenters, is a technology and a political issue. The political urge to inflict both kinds of spying is just that, a political one. As technology improves, it makes the job of spies and other criminals that much easier, increasing the spy's reach, and changing the character of society.

The technology is fascinating, but it's not divorced from society. We're all swimming in the same soup.

This is why the EU courts have a "right to be forgotten"

Possibly like the DNA database - really hard to get off. IIt's worse than just convicted criminals being stuck on the database though. A bunch of people are arrested, and have their photo and DNA taken and added to databases, but they're never tried and thus never convicted, but their details stay on the system unless they apply to have them removed.

Here is a case from 2008 where two people went to court to get their details taken off: http://news.bbc.co.uk/1/hi/uk/7764069.stm

And here's an earlier quote from a stupid judge suggesting everyone should be on the database: http://news.bbc.co.uk/1/hi/uk/6979138.stm

I look forward to privacy conscious individuals taking steps: https://cvdazzle.com/

This does not scale. It might work for a few people in a population of thousands, but beyond that false positives are going to be a real problem.

It's not going to be great for having 1 robbery and nabbing the next person your cameras ID. But in this day and age were the police won't go out for burglary calls we might end up with person A identified in X surveillance system incidents over Y years. That's an interesting step forward even if it isn't damning.

Is that really a step forward? When combined with the false positives issue, there could be a lot of harassment of innocents who accidentally fit a profile.

I work in facial recognition. When you say it does not scale, how so? I have a system with 975K people in it, the US registered sex offenders database, and with a single quad core laptop a lookup takes about 8 seconds. With a server class 32+ core system the lookup is nearly real time. How does that not scale?

I assume he more meant accuracy at scale - if you have a large population how does the accuracy do? Do you wind up with many close samples or are things pretty good?

Pretty much all FR systems generate a list of matches, ranked from closest match on down. The size of the list is configurable. It is also industry standard NOT to use the list as an authority. Given a list of high matches (greater than 90%) a human can quickly filter out obvious false positives, and then the remaining are retained for further consideration. It is also industry standard NOT to rely on FR alone; combining FR with other measures reduces the ability for a false positive.

The issue I find with FR is people expecting it to be some super technology, gleaming significant information from a dust spec. Its not like that. And the media is playing it up with unrealistic descriptions. Remember the original DOOM and it's quality of graphics? That is where FR is now. We've got a ways before the journalist hype is close to reality. And the mature technology will not be FR, but a comprehensive multiple biometric measuring system capturing far more than someone's face.

So, if you’re looking for 1% of a total population your false positive rate before human intervention is 99%. That's not useful for automated systems.

Further, identical twins are going to show an image that would fool human verification making this questionable for many tasks even with human supervision.

PS: Almost 1% of the population has an identical twin. (% of births is lower, but you get 2 twins.)

There is only one true match. So the system will always be generating false positives. This technology is not an authority, but a filter. Yes, identical twins will both be identified if they are both in the system and no additional biometric measures are included. Identical twins still have different retina, and due to lifestyles identical twins beyond age 30 can be distinguished apart fairly easily.

What's the "industry" you speak of? This article is suggesting this be used by minimum wage (or below[1]) workers in shops, bars and restaurants. How well do you think your industries norms will be adhered to?

[1] UK tried (is trying?) a "you must work for free in what would be a minimum wage job in a high street shop" scheme. http://www.bbc.co.uk/news/uk-politics-24742499

I think the OP was concerned about false positives, not processing times. If you ran, say, 10,000 faces through your system/day for a year, how many false positives would you get?

How does that reliably work? Because my impression from consumer products (say Picasa or Apple Photos) is, that starting with 35 to 50 tagged faces, the recommendations produce more false positives than correct ones. I thought that Google/Apple also used Maschine Learning techniques for face recognition.

Keep in mind: 1) there will always be more false positives because there is only one true match, and 2) consumer FR "toys" are nothing like an industrial FR application. It is night and day the difference between anything a consumer can play with and what police, corporations, and security agencies use.

The parent was talking about the statistics involved. Quoting lookup times is meaningless if those lookups aren't accurate. I could implement a system which always returns the first DB entry; lookups would take much less than 8 seconds, and by storing all other entries in /dev/null it has constant space usage ;)

This is clearly a play by the plastic surgery lobby to get people to have work done so they look less like known criminals.

The problem of false positives can and will simply be farmed out to a mechanical turk type of process ... a farm of low-cost workers will simply review hits, and if there's any possibility that it's a match, it will be escalated up several levels until action is taken.

If this was deployed in the US, I'd be concerned about being shot by the police because I was false-IDed as a violent criminal. Because in the US, it's become protect the police at all costs, including the cost of my life.

I wonder what the data protection implications of this are? I presume there are many. You're storing personal information on people, possibly transfering it to another businesses, analysing it and using it to make decisions....

    <span style="color: #FF2400">A</span>

Reply on news.ycombinator.com