Dubai detectives to get Google Glass
reuters.com
reuters.com
The idea of a wearable HUD is cool, but I don't see how any implementation can ever by trustworthy to non-wearers.
In effect, you seem to be opposed to replacing brain processing with computer processing, which is squarely on the wrong side of history.
This is going to happen, whether we like it or not. Banning it is a stopgap, we just need to start working on regulating it and defining the boundaries of such practices.
In keeping with traditional Islamic morality, both Federal and Emirate law prohibit homosexuality and cross-dressing with punishment ranging from long prison sentences, deportation, for foreigners, and the death penalty. (1)
Hence, it makes me want to throw out the idea of how effective this facial recognition really is. If your sole reason is marketing, you won't care much for effective use of said tech.
Quantity has a quality all of it's own. While currently the police can attempt to observe and memorize every face they see during their patrols, and where/when they saw it, doing this without a technological assist will be fruitless. Police without technology can memorize the faces of a few select persons of interests, but for tracking the entire public they are relatively worthless. This technology threatens to change this.
> "if they do _anything_ with it it will eventually mess up and they will be eviscerated in court."
Why are you so confident of that? UAE police have been credibly accused of systematic human rights violations in the recent past, and I've seen no hint of eviscerating over that...
The trouble is, so long as the cops have the ability to switch off the camera/pull the battery/cover the lens, we lose all the benefits of wearable video for police.
Before beating up Rodney King 2.0, they just flip the switch. Then in court, "The officers' cameras had a technical malfunction at the time and there is no video of the incident."
But if they want evidence of you doing/saying something incriminating, you bet your boots those cameras are rolling.
This is actually an interesting systems design problem, wherein the users are adversarial but in total control of the hardware.
(And yes, the wording is intended to resemble a common talking point about guns. The difference is that a mistake made with a camera is often correctable, a mistake made with a camera is often final).
1) Reverse burden of proof for charges against officers where footage is missing.
2) Tie payroll for the entire department to the availability of footage during a given pay period. Missing footage? Missing paychecks.
Without the video, the prosecution would have had eyewitnesses and the injuries to King, which were pretty damning. The defense would not have really had anything to counter that.
With the video, the defense was able to take a step by step approach to the defense. They showed that the first blows were in response to King being uncooperative or aggressive, and were probably justified. They then framed the issue as when, if ever, did the beating switch from justified force against an aggressive suspect and turn to excessive force against a man who had stopped fighting and submitted?
Each time King was hit, he'd twitch or kick or flail an arm or a leg. The defense started with the first blows, which they were able to justify because King did start out aggressive, and then they went blow by blow, looking at the position of each officer and what he could see, showing that each saw one of those kicks or flails of King's, and that from his position the officer could not see that was an involuntary response to a prior blow. From what the officer could see, King was still violently resisting, and so that officer would take a swing. The defense would show that caused an involuntary response that made the next officer think King was still fighting.
They went all the way through the video that way, getting the jury to focus on the difficulty of pointing to any point and saying that this was where the line was crossed.
This worked, and they got acquittal.
No video, and the jury would have likely focused on the totality of what happened to King, and then it would be hard not to find excessive force.
I think the video also made the prosecution overconfident. I think they thought it was going to as simple as showing up, playing the self-evident video, proving that the accused were the officers in the video and that the video was real, and that would be it.
Is the technology reliable to have a low false positive rate? Probably face recognition is good enough to classify the photograph of your friends in Facebook, and the friends of your friends. It has to choose a person in a set of ~100.
What happens when someone is falsely identified? Is the suspect released after a ID check or he is moved to the police station to a full check?
Can this be a proxy for racial profiling? There was a horror story about a face recognition software that didn't detect someone with black skin. This can have the reverse problem. http://www.youtube.com/watch?v=t4DT3tQqgRM
I'm guessing the same thing that happens if someone is falsely identified without Google Glass. Why would it be different?
For example, how will the system overcome image noise introduced by traditional dressing, such as the keffiyeh and hijab? I'm no expert, but have played around with haar face detection, and eigen/fisher face recognition. My experience was that factors like face position, facial adornments (caps, spectacles, moustaches, etc) and ambient lighting had significant effects on accuracy.
There was also the reported failure of face recognition in the Boston marathon bombing investigation [1].
[1]: http://nakedsecurity.sophos.com/2013/04/23/facial-recognitio...
This is a combination of Dubai's two greatest past times: spectacle and surveillance, so naturally they will implement this as quickly as possible, paying top dollar to contractors to make it work.
I'm sorry if this sounds jaded, but I lived this for 2.5 years, running projects whose sole purpose was pomp and circumstance for various Emiratis who have virtually unlimited (debt funded) budgets.
To all those saying it won't work because {{technical reason}} you're probably right, but it won't stop them from overpaying to do so.
Some more examples of this:
* They have a road toll system called Salik, which is like EZ Pass gone amuck. There are numerous tolls and each give the police the ability to track any car to a small area. So ther is very little auto crime.
* There are speed cameras about every 10km, which residents speed up and speed down to "obey" the speed limit. So to whomever was talking about high speed chases: they don't happen. They just wait for the criminal to stop, know exactly where he is, and then apprehend him.
* Most projects I worked on had a whole component about being "the best", "the biggest", "the tallest" where we would have to show why {{costly technology}} is the best in the world. There was no definition of "the best" beyond just claiming something.
* In my apartment building, there were cameras from outside my door to the carpark in the basement. And they were regularly watched. I know this because the door man would make it a habit of hitting the elevator button for me when I was about a meter away from the elevator, having seen me from CC cameras far below
* To those talking about how burqas and hijabs will make this technology impossible: most folks there don't wear traditional, national dress. 87% of the country are westerners who stay there for an average of 2 years and leave. Westerners and low-wage workers are who they want to track anyway.
* All internet is filtered by "du" and "Etisalat" (think ATT and Verizon) - by order of the state. But they don't maintain a common block list, so some websites work on du, some on Etisalat.
* There are large articles on this, but basically the UAE compelled Blackberry to allow them to install backdoors so that the government could circumvent secure communication (this was back when Blackberry was dominant).
And there are many more.
I'm not saying other countries don't have these problems. I'm more just listing these things out to show that this is exactly the kind of project that Dubai would love - high technology, spectacle, and above all controlling.
The fact that the door man would watch the feed and press the elevator button remotely was an unwanted and often reminder that I was constantly under surveillance even at my residence. It wasn't that I walked by the door man and he hit the button for me just as I walked up. It wasn't that he had a control for the elevator (which I get is common). It was that I was several floors above or below him and he was monitoring me and pressed the button for me. He even said once, when I asked him if he was the one doing, that "Yessir, I see you all the time."
It was creepy.
This is a state-level deal and FB has the most advanced face recognition software IIRC (I've read about this some months ago online), there's tons of $$$ to be made in the field.
Lamborghinis are fashion or do they have specific functions being part of police scheme of things?
High-speed police chases are a classic example of fiction over data. They do vastly more harm than they help, but we still have them because police officers grow up thinking they would chase bad guys in cars.
Dubai is surrounded by desert, where going 150+ in your $2 million car is something a lot of people would be tempted to do. If the police's top speed is 90, then the wealthy are above the law.
High-speed police chases do a lot of harm when they take place in dense urban/suburban neighborhoods. The use case for police Lamborghinis is on huge rural/desert highways.
Lamborghinis work best in video games though.
It is pure fashion.
It seems to me that putting that technology in the hands of everyone would make more sense than restricting it to a few.
Honestly, I think the product has been on borrowed time since the first person thought up the term "glasshole".
In ten years from now, everybody will probably wonder as well since by then, this technology (Glass + facial recognition) will most likely be mainstream.
But right now, it's prudent from Google to make a theoretical stand against that while Google Glass gains traction.
don't know why the downvotes - my phd research is in periocular recognition (using only the eye, eyebrow, and surrounding region to do verification and identification), but if you don't believe it is possible i suggest you check out some of the work in ieee-xplore