1) Give it some years to develop. Yes, I'm aware the accuracy right now is not 100%. Look at my other post here to get an idea of how the police will use a system like this, and you'll see where a system like this fits into their equation of work, and where a false alarm will be determined.
2) I know there are some really bad cops out there. Look at the link above and that fact that the system is automated. That means a computer will do the actual looking for faces, not a cop. Only when a possible match is found will a human actually see a live picture, and he/she will decide if it's a match or not. If it's a match, they still have to decide if it's a high enough priority to go after the person (ie, unpaid parking tickets vs. mass murderer vs. hostage situation already in progress elsewhere)
Now as for your other example, Person A commits a crime but looks like Person B, who just happens to also have a criminal record. That actually happens all the time, it's nothing new; this technology didn't create it, classic police work did. Oftentimes the police would think that Person B was actual suspect until they cross a piece of evidence that clears B of wrongdoing.
A good way to evaluate a system like this is to see how it would get used and how it fits into the system. Police wont rely on it 100% of the time, they'll continue to use their traditional methods of investigating to track down a wanted suspect. A system like this is not the one that throws people in jail, a case still has to be presented to the judge/jury and the decisions will be made from there.
btw - I myself have done some work in law enforcement. Chances are good it wont be cops manning workstations at HQ waiting for matches to appear on their screen. It'll be people like those operating 9-1-1 call centers. These people are not exactly tech savvy to know they're even able to abuse it. Yes, some can still abuse the system, but the system will be designed to minimize it, and abusing the system will actually make the person counter-productive.
And in a large city like NY or Tampa, these people would be so flooded with hits they wont have time to abuse the system, especially if cameras are outfitted in crime-ridden parts of town.
you make it seem like all cops do is abuse their power.
let me put it another way since i just thought of it...
all cops carry weapons like a handgun, pepper spray, and night sticks
when we walk down the street and happen to walk past a police officer, do we fear that they'll randomly pull their gun on us, randomly spray us with pepper, or randomly beat us with their sticks? The answer is no, we dont. If someone were to do that, that cop would permanently lose his badge.
Just because they have the tools and the tools pose a possibility of abuse doesnt mean that the tools will be abused. And even if that tool is abused, they will get caught, much like the Rodney King beating in LA many years ago.
ok, how many police officers are there in this country? Probably millions. How many have been caught abusing their power? About 100/year. 100 out of 1 million is less than 1%.
Less than 1%. You make it seem like every single cop in the country is gonna do nothing but abuse this thing. If they were to do that what would happen? All hell would break loose because all the cops stay at HQ sitting in front of CC'd TVs.
Think about it: COPS HAVE JOBS TO DO OTHER THAN ABUSIVELY TRACK INNOCENT PEOPLE!!
It's also important to note that those cops you talk about got caught. This means that those who abuse the system would sooner or later be caught themselves and other cops thinking about abusing it would learn from it.
Once again: Cops have a job to do. If they were to abuse it, they would hurt themselves and further injure the public by failing to catch truly wanted crooks. That would make them lousy cops.
My advice for you: Get a life, learn how to think for yourself, and learn how to draw your own intelligent conclusions.
note how i continually used the word 'suspect,' not 'convicted rapist'
This system is used to nap wanted suspects so that they can be brought before the judge and due process will take it's course from there. The FaceIt system does dictate that Person X goes to jail, a judge decides that. FaceIt is used to find Person X when we it knows what Person X looks like.
Also, look at my other post about how the police would actually use the system which also covers "what if two people look alike".
i should probably point out a major difference in 2001 from 1994.
The cameras being used for face recognition today are a lot better than those used 7 years ago for general surveillence. Also, the camera that made that shot of the boys leading the toddler out the mall was probably 3-4 years old already.
Point being that camera technology has come a long way since 1994.
ok, the guy would've stood a much greater chance at being nabbed. But think about the process the police would go through when an alert comes in...
-Alert arrives at HQ. Appearing on the screen are a picture of suspect, his/her crime(s), special notes from the police dept that wants that person, live picture of that person, location of the person.
-Analyst decides credibility and priority of the report. If a further look is deemed necessary, pass it off to an officer near the location of the suspect.
-Police stake out suspect. They have a photo of who they're looking for on their video screen inside their police car. They decide what to do about approaching the individual or not.
-If they approach and begin questioning the person right on the street, they can possibly determine whether not the alert was real or not. If they think the situation demands taking the guy in, then that's what it means.
As you can see, there are several spots where a false alarm would be tripped up. For example, if I'm 6'5" in height and have the same face as a 5'4" suspect, then the initial analyst looking at me on the street would see that I'm too tall to match the description of the suspect and mark the alert as a false alarm.
My point being is that the police will be relying on more than just the face an individual in deciding whether or not to take a person in for questioning. There are many other characterists to look for, like a tatoo on the ankle or a scar on the arm. Prior to automated face recognition, this is how things have been done, and this is how I think it'll happen in the future when this stuff is widely used.
in addition to that, make it so that police depts react when alarms go off when a suspect is identified on the street. It appears here that was the problem.
Cameras are one thing, getting police to act is another.
... is why this is such a big issue. I would prefer wanted criminals be caught through a technique like this. They're dangerous to our society and dont belong on the streets.
I know a lot of people are worried that a system like this can be abused by authorities to track people. I have two uncles that are former police officers (one now is in the Secret Service, other died). Let me explain the point of view of the current SS agent:
There is so much work that a police dept in a major city like NY or Tampa that has to be done that there is no room to abuse a system like an automated facial recognizer. If someone were to abuse it, his/her overall job performance would go down because they would be tracking innocent people instead of catching wanted suspects.
I also have an example of a situation where this would work. I live in Philadelphia. About 2 years there was a serial murderer and rapist in Center City, and got dubbed the name Center City Rapist. A picture of the guy was found and wanted signs appeared all over town, on lampposts, park benches, etc. Also on those signs were how he attacks and how he targets single women who live alone. But the guy got away.
Last month his DNA was found on a rape & murder victim in Denver, Colorado.
If FaceIt were running on Denver and have the Center City Rapist's photo in the db, that guy would have been caught because of his high profile from Philly and perhaps one young woman would still be alive today because of FaceIt.
By doing this, they are saving their taxpayers a bundle (easily billions) by not spending so much on licensing fees from Sun or Microsoft. That money saved can go to greater things like making better schools, etc.
Perhaps congress should visit our German buddies and see how a switch to OSS can benefit the American public. A little bit of seeing what's happening abroad could go a long way.
What's your opinion of organizations providing funds to political campaigns in exchange for laws/policies/etc that benefit the organization? Could this be considered bribing on behalf of the funding organization and accepting a bribe by 'returning the favor?' If not bribes, would you consider this practice ethical?
I ask this question in how it pertains to the situation of organizations with deep pockets such as the RIAA funding lawmakers to create laws like the DMCA and other laws that are currently coming down the pike.
Also, what advice would you give to shallow-pocket organizations such as the EFF or EPIC in fighting to keep the rights of honest, well meaning Internet users?
It wasn't the impact that took the towers down, it was the fire that followed (which is why hydrogen as jet fuel is being discussed).
The towers were able to withstand the impact of the jets just fine, as they were designed to. It was the burning of the fuel that softened the support columns that ran vertically through the building. When those support columns became too soft, they fold. The sheer weight of the building above the soft spot caused the collapse.
That's why Tower 2 fell first - the plane hit near the 70-80th floor or something. Tower 1 was hit at the 103rd, meaning the columns there had to soften up a lot more than Tower 2's columns before they collapsed.
The idea of using hydrogen as jet fuel is that it burns at a much lower temperature. Hence if a hydrogen powered jet were to strike a sky scraper, firefighters would have a large amount of time to put the fire out and engineers time to build up support to save the building as a whole.
One place this did happen was (IIRC) in the 1930s. During heavy fog, a twin engine biplane rammed the Empire State Building. Granted it's a much smaller plane, but it still caused a fire, just not as large as a few weeks ago.
Windows XP may have fared poorly because it's brand new and so is it's API. On the other hand, Linux and Win2K have had a few years in which their APIs have matured/improved and some of their kinks or other bottlenecks straightened out. Point being that the XP API may still have those kinks in their code, causing some slowness compared to other systems.
Of course, MS could be simply re-using the 2K API in XP, something they should REALLY take a second look at doing given these results.
As it is now on the OS X system, Aqua is very limited in preferences like color schemes and such when compared to other GUI systems like Windows or KDE. By having it elsewhere, it opens the door for more customization. If others like that ability, it may encourage Apple to add such features to Aqua.
At least I hope so. The link above appears slashdotted and I cant see it (yet). I hope Apple left the door open for those kind of customizations.
where did the author see the note about the Win Media Player being able to shutoff other software? Looked through the eWeek article, didnt see it.
Even still, such a thing as arbitary software shutoff would of course be illegal.
Analogy: I own a Ford F1 pickup and a Saturn. Let's say I use the Ford to go to the furniture store to pickup my new couch. GM can't shutoff my Ford because I didnt use the Saturn. Such an act would be anti-competitive. Doesnt matter to GM that the couch wont fit in the Saturn.
Translation: I have a RealVideo clip I wish to view. If MS were to have WMP arbitrarily shutoff RealPlayer, it would be anticompetitive because people wont be able to use RP over WMP. Doesnt matter to MS that WMP doesnt support RealVideo.
Of course, this is MS, who's already been found an illegal monopoly etc etc (twice!). Lastly, IANAL.
(ot: I am reading a book right now on the whole antitrust case. I'm planning on a/. Book Review when I get done. The last chapter is titled 'MS Loses Even If It Wins' Book is about 400 pages, and I'm about 250 through it. Should be interesting...)
yeah, i agree, who follows licenses? Well, if this were to go to court or something, the licenses would be brought up by the lawyers.
Also, corporate america follows the EULA licenses. First day on the job you'll get warned by them on using software outside the licenses. If a company were to be caught, then that's bad press for that company.
Also, we/.ers seem to get all up in arms when a company sells GPL software w/o source code, so I think there are people out there who do follow them and want them followed.
How did you figure out that the product was full of GPL code and such? From the looks of things, it appears you had to reverse engineer binary code and hack out a root passwd. IANAL, but chances are good xStore put in the license agreement that you werent allowed to do those two things. You may run into trouble with that should everything turn out legit. Yes, they may have breached the GPL, but their agreement probably restricted you from those activities.
Basically, it's an issue of risk. If it turns out that they have no GPL violations, then you could get nailed for breaking the license they provided. On the other hand, you could show that they broke the GPL prior to specifying the license terms you use the product with, either voiding their license or something of that nature.
you're probably already aware of this, but swing has not been out 5 years yet. I was only introduced in (i think) 98. Overall, however, I argee with what your point is.
They're is doing research along this same theme. They someday want robots to be able to be controlled either from the ground or the space station by a person wearing a VR suit. The human can see a 3d image of what the robot is working on and they're working on having the bot transmit some kind of feel (a sense of touch) back to the human. Advantage to this is the ability to make the robot fit into small or dangerous areas and not subjecting a live person to danger.
Slashdot Mirror
1) Give it some years to develop. Yes, I'm aware the accuracy right now is not 100%. Look at my other post here to get an idea of how the police will use a system like this, and you'll see where a system like this fits into their equation of work, and where a false alarm will be determined.
2) I know there are some really bad cops out there. Look at the link above and that fact that the system is automated. That means a computer will do the actual looking for faces, not a cop. Only when a possible match is found will a human actually see a live picture, and he/she will decide if it's a match or not. If it's a match, they still have to decide if it's a high enough priority to go after the person (ie, unpaid parking tickets vs. mass murderer vs. hostage situation already in progress elsewhere)
Now as for your other example, Person A commits a crime but looks like Person B, who just happens to also have a criminal record. That actually happens all the time, it's nothing new; this technology didn't create it, classic police work did. Oftentimes the police would think that Person B was actual suspect until they cross a piece of evidence that clears B of wrongdoing.
A good way to evaluate a system like this is to see how it would get used and how it fits into the system. Police wont rely on it 100% of the time, they'll continue to use their traditional methods of investigating to track down a wanted suspect. A system like this is not the one that throws people in jail, a case still has to be presented to the judge/jury and the decisions will be made from there.
btw - I myself have done some work in law enforcement. Chances are good it wont be cops manning workstations at HQ waiting for matches to appear on their screen. It'll be people like those operating 9-1-1 call centers. These people are not exactly tech savvy to know they're even able to abuse it. Yes, some can still abuse the system, but the system will be designed to minimize it, and abusing the system will actually make the person counter-productive.
And in a large city like NY or Tampa, these people would be so flooded with hits they wont have time to abuse the system, especially if cameras are outfitted in crime-ridden parts of town.
you make it seem like all cops do is abuse their power.
let me put it another way since i just thought of it...
all cops carry weapons like a handgun, pepper spray, and night sticks
when we walk down the street and happen to walk past a police officer, do we fear that they'll randomly pull their gun on us, randomly spray us with pepper, or randomly beat us with their sticks? The answer is no, we dont. If someone were to do that, that cop would permanently lose his badge.
Just because they have the tools and the tools pose a possibility of abuse doesnt mean that the tools will be abused. And even if that tool is abused, they will get caught, much like the Rodney King beating in LA many years ago.
ok, how many police officers are there in this country? Probably millions. How many have been caught abusing their power? About 100/year. 100 out of 1 million is less than 1%.
Less than 1%. You make it seem like every single cop in the country is gonna do nothing but abuse this thing. If they were to do that what would happen? All hell would break loose because all the cops stay at HQ sitting in front of CC'd TVs.
Think about it: COPS HAVE JOBS TO DO OTHER THAN ABUSIVELY TRACK INNOCENT PEOPLE!!
It's also important to note that those cops you talk about got caught. This means that those who abuse the system would sooner or later be caught themselves and other cops thinking about abusing it would learn from it.
Once again: Cops have a job to do. If they were to abuse it, they would hurt themselves and further injure the public by failing to catch truly wanted crooks. That would make them lousy cops.
My advice for you: Get a life, learn how to think for yourself, and learn how to draw your own intelligent conclusions.
note how i continually used the word 'suspect,' not 'convicted rapist'
This system is used to nap wanted suspects so that they can be brought before the judge and due process will take it's course from there. The FaceIt system does dictate that Person X goes to jail, a judge decides that. FaceIt is used to find Person X when we it knows what Person X looks like.
Also, look at my other post about how the police would actually use the system which also covers "what if two people look alike".
Lastly, learn to think for yourself.
i should probably point out a major difference in 2001 from 1994.
The cameras being used for face recognition today are a lot better than those used 7 years ago for general surveillence. Also, the camera that made that shot of the boys leading the toddler out the mall was probably 3-4 years old already.
Point being that camera technology has come a long way since 1994.
ok, the guy would've stood a much greater chance at being nabbed. But think about the process the police would go through when an alert comes in...
-Alert arrives at HQ. Appearing on the screen are a picture of suspect, his/her crime(s), special notes from the police dept that wants that person, live picture of that person, location of the person.
-Analyst decides credibility and priority of the report. If a further look is deemed necessary, pass it off to an officer near the location of the suspect.
-Police stake out suspect. They have a photo of who they're looking for on their video screen inside their police car. They decide what to do about approaching the individual or not.
-If they approach and begin questioning the person right on the street, they can possibly determine whether not the alert was real or not. If they think the situation demands taking the guy in, then that's what it means.
As you can see, there are several spots where a false alarm would be tripped up. For example, if I'm 6'5" in height and have the same face as a 5'4" suspect, then the initial analyst looking at me on the street would see that I'm too tall to match the description of the suspect and mark the alert as a false alarm.
My point being is that the police will be relying on more than just the face an individual in deciding whether or not to take a person in for questioning. There are many other characterists to look for, like a tatoo on the ankle or a scar on the arm. Prior to automated face recognition, this is how things have been done, and this is how I think it'll happen in the future when this stuff is widely used.
in addition to that, make it so that police depts react when alarms go off when a suspect is identified on the street. It appears here that was the problem.
Cameras are one thing, getting police to act is another.
... is why this is such a big issue. I would prefer wanted criminals be caught through a technique like this. They're dangerous to our society and dont belong on the streets.
I know a lot of people are worried that a system like this can be abused by authorities to track people. I have two uncles that are former police officers (one now is in the Secret Service, other died). Let me explain the point of view of the current SS agent:
There is so much work that a police dept in a major city like NY or Tampa that has to be done that there is no room to abuse a system like an automated facial recognizer. If someone were to abuse it, his/her overall job performance would go down because they would be tracking innocent people instead of catching wanted suspects.
I also have an example of a situation where this would work. I live in Philadelphia. About 2 years there was a serial murderer and rapist in Center City, and got dubbed the name Center City Rapist. A picture of the guy was found and wanted signs appeared all over town, on lampposts, park benches, etc. Also on those signs were how he attacks and how he targets single women who live alone. But the guy got away.
Last month his DNA was found on a rape & murder victim in Denver, Colorado.
If FaceIt were running on Denver and have the Center City Rapist's photo in the db, that guy would have been caught because of his high profile from Philly and perhaps one young woman would still be alive today because of FaceIt.
The murderer and rapist is still on the run.
By doing this, they are saving their taxpayers a bundle (easily billions) by not spending so much on licensing fees from Sun or Microsoft. That money saved can go to greater things like making better schools, etc.
The article starts out saying that Congress wants to outlaw Open source via the SSSCA.
Perhaps congress should visit our German buddies and see how a switch to OSS can benefit the American public. A little bit of seeing what's happening abroad could go a long way.
Mr Reed:
What's your opinion of organizations providing funds to political campaigns in exchange for laws/policies/etc that benefit the organization? Could this be considered bribing on behalf of the funding organization and accepting a bribe by 'returning the favor?' If not bribes, would you consider this practice ethical?
I ask this question in how it pertains to the situation of organizations with deep pockets such as the RIAA funding lawmakers to create laws like the DMCA and other laws that are currently coming down the pike.
Also, what advice would you give to shallow-pocket organizations such as the EFF or EPIC in fighting to keep the rights of honest, well meaning Internet users?
I think the keyword here is 'Attempt'
What kind of privacy isues were the FTC trying to push? I know right-to-know how personal data would be used was one, but what else?
hey, I said IIRC. Apparently I didnt RC!
It wasn't the impact that took the towers down, it was the fire that followed (which is why hydrogen as jet fuel is being discussed).
The towers were able to withstand the impact of the jets just fine, as they were designed to. It was the burning of the fuel that softened the support columns that ran vertically through the building. When those support columns became too soft, they fold. The sheer weight of the building above the soft spot caused the collapse.
That's why Tower 2 fell first - the plane hit near the 70-80th floor or something. Tower 1 was hit at the 103rd, meaning the columns there had to soften up a lot more than Tower 2's columns before they collapsed.
The idea of using hydrogen as jet fuel is that it burns at a much lower temperature. Hence if a hydrogen powered jet were to strike a sky scraper, firefighters would have a large amount of time to put the fire out and engineers time to build up support to save the building as a whole.
One place this did happen was (IIRC) in the 1930s. During heavy fog, a twin engine biplane rammed the Empire State Building. Granted it's a much smaller plane, but it still caused a fire, just not as large as a few weeks ago.
Windows XP may have fared poorly because it's brand new and so is it's API. On the other hand, Linux and Win2K have had a few years in which their APIs have matured/improved and some of their kinks or other bottlenecks straightened out. Point being that the XP API may still have those kinks in their code, causing some slowness compared to other systems.
Of course, MS could be simply re-using the 2K API in XP, something they should REALLY take a second look at doing given these results.
As it is now on the OS X system, Aqua is very limited in preferences like color schemes and such when compared to other GUI systems like Windows or KDE. By having it elsewhere, it opens the door for more customization. If others like that ability, it may encourage Apple to add such features to Aqua.
At least I hope so. The link above appears slashdotted and I cant see it (yet). I hope Apple left the door open for those kind of customizations.
where did the author see the note about the Win Media Player being able to shutoff other software? Looked through the eWeek article, didnt see it.
/. Book Review when I get done. The last chapter is titled 'MS Loses Even If It Wins' Book is about 400 pages, and I'm about 250 through it. Should be interesting...)
Even still, such a thing as arbitary software shutoff would of course be illegal.
Analogy: I own a Ford F1 pickup and a Saturn. Let's say I use the Ford to go to the furniture store to pickup my new couch. GM can't shutoff my Ford because I didnt use the Saturn. Such an act would be anti-competitive. Doesnt matter to GM that the couch wont fit in the Saturn.
Translation: I have a RealVideo clip I wish to view. If MS were to have WMP arbitrarily shutoff RealPlayer, it would be anticompetitive because people wont be able to use RP over WMP. Doesnt matter to MS that WMP doesnt support RealVideo.
Of course, this is MS, who's already been found an illegal monopoly etc etc (twice!). Lastly, IANAL.
(ot: I am reading a book right now on the whole antitrust case. I'm planning on a
The ZDNet article makes a reference to 'Uncle Billy'. Remindes me of Uncle Sam....
The Northwestern site appears to be slashdotted.... anybody got the relevant text from that pdf file?
THATS the term i was looking for earlier!!! Catch 22!! That's the risk the talked about but couldnt think of the right phrase!
yeah, i agree, who follows licenses? Well, if this were to go to court or something, the licenses would be brought up by the lawyers.
/.ers seem to get all up in arms when a company sells GPL software w/o source code, so I think there are people out there who do follow them and want them followed.
Also, corporate america follows the EULA licenses. First day on the job you'll get warned by them on using software outside the licenses. If a company were to be caught, then that's bad press for that company.
Also, we
your right, that's why i point out that it's a risky move trying to force this issue.
How did you figure out that the product was full of GPL code and such? From the looks of things, it appears you had to reverse engineer binary code and hack out a root passwd. IANAL, but chances are good xStore put in the license agreement that you werent allowed to do those two things. You may run into trouble with that should everything turn out legit. Yes, they may have breached the GPL, but their agreement probably restricted you from those activities.
Basically, it's an issue of risk. If it turns out that they have no GPL violations, then you could get nailed for breaking the license they provided. On the other hand, you could show that they broke the GPL prior to specifying the license terms you use the product with, either voiding their license or something of that nature.
you're probably already aware of this, but swing has not been out 5 years yet. I was only introduced in (i think) 98. Overall, however, I argee with what your point is.
This surgery should be encouraging news for NASA.
They're is doing research along this same theme. They someday want robots to be able to be controlled either from the ground or the space station by a person wearing a VR suit. The human can see a 3d image of what the robot is working on and they're working on having the bot transmit some kind of feel (a sense of touch) back to the human. Advantage to this is the ability to make the robot fit into small or dangerous areas and not subjecting a live person to danger.