> what if police officers were equipped with body cameras that could alert them the instant they come into view of a handgun, or some other weapon?

And then

> Overall, the optimized model achieved 84% accuracy, with zero false negatives [...] and 57 false positives.

So in the what-if scenario above, 16% of cases would falsely escalate a situation dramatically? Seems like a horrible idea. And "really good" seems like a stretch.

16% false positive is hilariously bad, especially given how rare actual handguns will be.

This headline shouldn't be tolerated here.

On the other hand missing completely it in 3 out of 30 tests (samples with a number of selected "optimizations") is a "terrible" 10% is awfully worse (in an alarm system).

>For the experiments, we selected 30 scenes from the previously used videos with the next requirements. Each scene is made up of at least 5 frames, filmed in a fixed scenario, i.e., in the same place, and the pistols are clearly visible to a human viewer. ... The model successfully detects the pistol in 27 scenes with an average time interval AATpI=0.2 seconds, which is good enough for an alarm system. The detector fails to detect pistols only in three scenes.

I am probably particularly picky but (from the paper):

>This section analyzes the performance of our best detection model on seven public videos, with low qualities, downloaded from youtube. Six of them are pieces of well known films from the 90s, James Bon: The World is Not Enough (video 1), 3 pieces from Pulp Fiction (video 2, 3 and 4), Impossible Mission: Rogue Nation (video 5) and Mister Bin (video 6). The seventh video is a long pistol threatening video (video 7).

... if you cannot spell Bond or Mister Bean, and you classify Mission Impossible: Rogue Nation as a 90's movie (calling it Impossible Mission) you don't look like having been particularly accurate when writing the article, if the same accurateness is in the actual numbers, there is a problem.

Anyway the "final" sample (30 scenes) is so small that it makes very little sense to derive anything from it.

What happens when it pushes the panic button when a firearm is present, but carried without threat, legally?

That sort of accuracy seems like about three orders of magnitude off from what would be useful in the field.

Unless you just basically want a drug-sniffing dog, for guns. That is, something to "scienceify" something that would otherwise be problematic (searching this car, or shooting that suspect).

"Hey judge, the DOG said it, it must be true!"

"Hey judge, the COMPUTER said it, it must be true!"

That number should be compared to the rate human officers unnecessarily escalate a situation because they think there's a gun. I don't know that number but it could easily be higher than 16%.

It should really be compared to the number of instances human officer's act as if there could be a gun " just in case ". Like in a traffic stop.

If officer's don't have to potentially gun someone down because they made a hectic movement 84% of the time, that's a pretty big win.

Imagine 84% of " black guy mysteriously died" actually led to convictions of officers because their ai thing told them "definitely no gun, bro"

> Imagine 84% of " black guy mysteriously died" actually led to convictions of officers because their ai thing told them "definitely no gun, bro"

I imagine that, like body cameras, the AI thing will, in those events, conveniently be deactivated, malfunctioning, misplaced, or have it's records lost.

probably. but you can only pull that stunt so many times.

> probably. but you can only pull that stunt so many times.

How many? Because apparently we haven't yet reached it with body cams.

Until you've produced a sample. Then some statistician can write a paper on it and someone takes it to the supreme court and 20 years later they will decide that there was probably something fishy going on.

Like I know?

But if you compound different systems, like the ai handgun thing and dont store the data in the same place, it becomes exponentially less likely that suddenly 2 critical systems failed at the same important moment for some reason.

the real issue with police brutality is the fact that it is, in fact, legal, and police is trained to respond in such a way. what protesters get wrong about police brutality is that its supposedly illegal. it is not. most instances of police killing someone are actually legal. which doesn't make it right. but its legal. saying somethings illegal when its not just makes you look like an idiot.

people need to protest for better laws, not getting individual cops sentenced, because that wont ever happen with any kind of regularity.

It could easily be lower than 16% too. Much lower.

Right, I'm saying I don't know the number, but this algorithm shouldn't be judged against a theoretical perfect officer, because that's not reality.

There are also a lot of legal gun owners who have a gun but are not a threat.

They seem to be spotting visible guns, not concealed ones, and the focus is on guns being brandished instead of holstered.

This is preliminary work but it wouldn't be useful to have an alarm go off every time another police officer or security guard comes into view with a gun on his hip.

That's true in the US, but not in many other places (the UK for example).

I guess you'd have to compare the failure rate to that of humans, similar to automated driving.

Beside that, the camera would have to view the subject from the correct angle, so, from the back, most likely position if they're carrying concealed.

The usefulness of a system like this would be where the police have a suspect cornered with their weapons drawn, but the suspect refuses to comply with instructions. If the suspect makes a suspicious move (e.g. as if he is reaching to pull a gun from a back pocket) the police will shoot immediately, since they would not have time to react if they waited to see if he was pulling a gun or not. Right now, the human system provides a 100% false positive rate. A weapon system that incorporates this detector would allow the police to pull the trigger when the suspect makes the suspicious move, but delay the firing of the weapon until the system detects a gun. Even with a 16% false positive rate it would save lives.

That said, I don't think police would want it if it might increase the risk to their lives.

Adding any kind of electronic firing control mechanism to a service pistol is a non-starter for a bunch of reasons:

MTBF: Even small guns produce recoil, which over time will damage or degrade any electronics built into the firing mechanism.

Security: if a "computer" of whatever complexity is controlling the firing of the gun it presents surface area both for attacks and bugs.

Rank and file acceptance: No officer will want to go into the field with a service weapon fitted with "safety restrictions" or "casualty reduction mechanisms" other than the trigger safety.

I agree with everything you said, except MTBF.

The training guns could be separate. Surely the "on-duty" guns are not being fired often enough that this will be an issue at all... right? Right??

#1 feeds into #3-- How confident are your rank and file going to be when the "training" guns fail regularly. Also there is something to be said for training with the exact same weapon you plan to use in the field.

"But what if police officers were equipped with body cameras that could alert them the instant they come into view of a handgun, or some other weapon? Or if bank’s security system could sound an alarm the instant cameras detect a robbery in progress?"

Why are we even bothering with spending money trying to create technology like this? How often do banks actually get robbed anymore by gun-wielding maniacs? And how often do officers actually die in the line of duty by getting shot by armed citizens? This is the more frequent case than the bank robbery, and even that is exceedingly rare. Only ~140 officers died on the job in 2016 and of those only 64 died from being shot. Factor in that even in a fair amount of circumstances when they know someone is armed they still end up getting shot and that number of preventable deaths (in theory) gets even lower.

What I always find strange is the obsession with "officer safety". Being a cop isn't even one of the top 10 most dangerous jobs in the country, statistically speaking. It's far more dangerous to be in logging, farming, construction, fishing, transportation, etc. And that's comparing the numbers for a very high year like 2016. In 2013 the number of officer deaths was less than half of 2016 - only 31 cops died from being shot - out of ~900,000 officers in the nation.

This kind of thing is just a colossal waste of money.

"Systems like these still have a ways to go, especially since those false alarms could have deadly consequences for civilians being approached by body camera-wielding cops."

That's the biggest issue I have with this. This will become a new reason for them to detain and probably stop and frisk. "The device alerted me to a possible threat" - this will be the excuse and this will lead to reduced rights because the courts will, of course, consider it "reasonable suspicion" based on the technology. There will also be no incentive to increase accuracy since interacting with more citizens makes/catches more criminals, which means numerically they are doing a better job.

> What I always find strange is the obsession with "officer safety". Being a cop isn't even one of the top 10 most dangerous jobs in the country, statistically speaking. It's far more dangerous to be in logging, farming, construction, fishing, transportation, etc.

I don't think it has to do with the numbers exactly. If you are a skilled logger and don't make any mistakes, you are likely pretty safe. As a skilled cop, without any mistakes, your life can still be in jeopardy at the drop of a hat. No amount of safety gear can protect a cop from someone pulling a knife or other weapon with intent to harm either the cop or an innocent civilian.

I think there is also the aspect that cops are public servants. It's compassion that makes me want to keep these folks alive who are voluntarily taking risks to protect me and my community.

I think the numbers are also not totally representative - are all 'cops' lumped in together for the statistic about job safety? Does that include detectives doing mostly office or investigative work? Does it include officers writing parking tickets? Does it blend cops patrolling the suburbs with cops patrolling inner cities?

I don't have any sources about my claims here, I don't really know a lot about the dangers of logging or the statistics about the dangers of being a cop.

---

What bothers me the most about this is that visual presence of a firearm is immediately associated with criminal behavior or intent.

That's a good point about the risk not being in control of the officer through something like skill. Though, I don't think that changes the essence of my point, which could be summed up as diminishing returns. If we had a much higher rate of officer fatalities, then I would say pursuing this technology makes sense. But I think the resources can be allocated better elsewhere.

Totally agree that the mere presence of a gun being an alarm bell of sorts is the biggest issue here.

I disagree, I think the presence of a gun is definitely cause for alarm. The only reason anyone should be carrying a gun that is visible to others is if they intend to use it. If you want to carry one regularly (as opposed to only when you're going to use it), get a concealed carry permit, and keep it concealed.

I strongly disagree that having a weapon on display is only done if there is intent to use it. In some areas of the south, open carry is quite common. It should also be noted that getting a concealed carry permit is not always that easy, depending on where you live.

The reason I have no issue with open carry is that it 1) makes the weapon more quickly accessible should it be needed and 2) it acts as a deterrent. I think having a gun visible to any possible violent criminals would make them far less likely to choose that individual than one who may or may not be concealing it in their pant leg, purse, backpack, etc.

In fact, those two reasons are exactly why cops open carry when on duty. They need to have quick access just in case shit goes down and it makes people think twice (and then a third and fourth time) if they see an armed officer - how likely is it that someone is going to try to assault a cop and escape arrest if they could be shot in ~1 second?

Police by nature cannot be everywhere at all times. Even in dense cities it is not realistic to expect an officer to be able to respond within 30 seconds - a length of time that means the world when it comes to violent crime, whether it involves a gun or not. The further out you go from the dense cities the longer the response time, making it even more critical to have protection for yourself. This is the entire reason why the second amendment exists - to be able to protect yourself when the government cannot enforce and defend your rights for you.

> it acts as a deterrent

I'm pretty sure this has been thoroughly debunked. The only thing that carrying a weapon does is make it more likely that someone will die if an incident occurs. Not only because now there's a gun involved, but also because if the criminal's victims are armed with guns, then the criminal is going to be armed with a gun too.

> In fact, those two reasons are exactly why cops open carry when on duty.

What cops do and don't do is pretty much irrelevant to talking about the general public. They actually have the legal authority to do things like shoot people, or just to arrest them in general, meaning the presence of a cop (armed or otherwise) acts as a deterrent to crime already.

> This is the entire reason why the second amendment exists - to be able to protect yourself when the government cannot enforce and defend your rights for you.

No, the second amendment exists because it dates to when militias were common (in fact it explicitly mentions militias). I believe it was also meant as an escape hatch to allow another revolution if the US government got out of hand (which comes back to the whole militia thing). It was not intended as a deterrent for ordinary crime.

That has not been debunked - I'd say the jury is still out, at best. Many of the studies that supposedly debunk this claim have issues like including suicides, cherry picking specific violent crimes and excluding others, not accounting for gun purchases made in high risk areas because people buy out of fear (meaning violent crime was on the rise before the purchases), etc. One of the most reasonable studies I've seen is the one conducted by the CDC and was commissioned by the Obama administration.

http://www.cnsnews.com/news/article/cdc-study-use-firearms-s...

The CDC concluded that it is an important deterrent to crime and found consistently lower injury rates for armed citizens compared with victims who used other defense methods (or had no defense method at all).

What cops do is absolutely not irrelevant in this regard. I would like to remind you that ordinary citizens can usually legally detain and physically restrain a criminal and can perform a citizen's arrest when they witness a felony crime in progress. What better way to do that than with a gun? Also, any person can legally apply lethal force if their life is in jeopardy. Cops are granted the extra benefit of being able to apply lethal force when someone else's life is in jeopardy and theirs is not.

Yes, a crucial reason behind the second amendment was to maintain the ability to fight against a tyrannical government. Another reason is the ability to fight on behalf of your government on an as-needed, short-notice type of scenario. The third main reason is what I mentioned - self-defense. The threats of that time were much greater - angry Brits that we snubbed, Native Americans that we killed, wild animals, etc. Plus the need to hunt. Being armed was a necessity. It was so obvious, that I think that part didn't even get mentioned in the second amendment. However, later on a number of states added specific state constitutional provisions that secured the right to "bear arms in defense of themselves and the state". These provisions usually included an element of self/personal defense as well as the state or "common defense". These are separate things. A very clear example is Colorado - "The right of no person to keep and bear arms in defense of his home, person and property, or in aid of the civil power when thereto legally summoned, shall be called in question; but nothing herein contained shall be construed to justify the practice of carrying concealed weapons."

Please also keep in mind that since the constitution trumps all other laws (federal, state, etc.), all other laws that are passed must conform to the constitution to be considered legal and valid. Colorado's language is extremely clear and plain. It has never been overturned or deemed unconstitutional. This therefore means that the second amendment must contain the right to personal self-defense not in service of the nation. And it should be clear based on these state provisions that were all passed that it was clearly a major element.

> This therefore means that the second amendment must contain the right to personal self-defense not in service of the nation.

No, it just means the second amendment doesn't prohibit that.

Re false positives...

God I can already imagine it! "Drop the gun! Drop the gun! Pop pop pop"

Can't drop what you aren't holding.

That's a great point, what a terrible situation.

Am I misreading this article. They had 30 test scenes they tested this with, and in those 30 tests it returned 57 false positives? That seems super useless. It sounds like it is basically just throwing alerts for every scene.

The "scenes" are likely hours of video footage, but yes this is certainly missing a denominator. I think the best unit is probably "fraction of video time a handgun is on screen" but that's not told in the article.

I live in the south, it is perfectly legal and socially acceptable to carry a handgun, concealed or open. Something like this could add context to a situation, but sounding an alarm because of a handgun present sounds like a bit much.

So 3d print modifications to your handgun that alter its shape but not its functionality. Not a difficult workaround.

A lot of common security measures can be defeated with modest technical skill and diligence. Fortunately for the police and the public, the vast majority of armed criminals are not even modestly technically skilled and diligent. They're largely fools with poor impulse control. If you are smart enough to spot weaknesses in security measures and implement workarounds, you probably already have a job with better long term prospects than "bank stickup guy."

There are some criminals that get rich from armed robbery, but your typical armed bank robber has a short and none-too-lucrative career:

http://www.businessinsider.com/people-are-robbing-banks-in-r...

https://www.thestar.com/news/insight/2013/04/26/bank_robbers...

https://www.theguardian.com/commentisfree/2014/jul/04/where-...

> A lot of common security measures can be defeated with modest technical skill and diligence. Fortunately for the police and the public, the vast majority of armed criminals are not even modestly technically skilled and diligent.

Yes - although almost none of them have the technical expertise to manufacture a gun and yet significant number of criminals own them. Similarly criminals don't need the expertise to own a "gun obfuscating device" to own one.

There's a large, legal market for firearms in the US, so it's not hard for a criminal to take advantage of that. What's going to produce a large, legal market for dedicated gun obfuscating devices that criminals can similarly take advantage of? I don't think that there is going to be one. I think it will be more like a criminal getting access to e.g. plastic explosives: either he's part of a criminal organization, already has a specialized day job that allows for diversion, or has reasonably high technical skill and can manufacture the materials in a home workshop. I'm not sure the "diversion from legitimate business" path even works here, because again, I don't see the business need for disguising open carried guns. If Johnny Hothead, having no specialized connections, gets an offer to sell him some explosives then it's probably coming from an agent of law enforcement rather than the local Crime Depot.

simply shroud your firearm in a piece of cloth like the hitmen in gangster movies.

Or a piece of paper?

It seems something like this would be great for airport scanners too, although the biggest benefit would be in luggage scanning. I think the challenge would be finding a diverse enough set of true positives to train the model.

I cannot believe no one had pointed out the matching Robocop scene yet: https://www.youtube.com/watch?v=_mqDjcGgE5I scnr