Slashdot Mirror
Shadow Analysis Could Spot Terrorists
Hugh Pickens writes "An engineer at Jet Propulsion Labs says it should be possible to identify people from the way they walk — a technique called gait analysis, whose power lies in the fact that a person's walking style is very hard to disguise. Adrian Stoica has written software that recognizes human movement in aerial and satellite video footage by isolating moving shadows and using data on the time of day and the camera angle to correct shadows that are elongated or foreshortened. In tests on footage shot from the sixth floor of a building, Stoica says his software was indeed able to extract useful gait data. Extending the idea to satellites could prove trickier, though. Space imaging expert Bhupendra Jasani at King's College London says geostationary satellites simply don't have the resolution to provide useful detail. 'I find it hard to believe they could apply this technique from space,' says Jasani." Comments on the article speculate on the maximum resolution possible from KH-11 and KH-12 spy satellites.
Go ahead, develop more technology, there's always around it.
Who puts a spysat in geostationary orbit? It's way too high, you'd need a telescope that dwarfs Hubble to get a decent view. You put spysats in the lowest orbit you can get away with, and you make sure that you have enough of them that any target of interest will be covered frequently enough for your purposes.
Real Daleks don't climb stairs - they level the building.
An engineer at Jet Propulsion Labs says it should be possible to identify people from the way they walk -- a technique called gait analysis, whose power lies in the fact that a person's walking style is very hard to disguise.
I knew it! The Ministry of Silly Walks is really just a subdivision of MI6!
Defeated by a simple 2 inch lift in one shoe.
Obviously, but this isn't exactly rocket science.
Just point at the screen and say "Enhance this part!" There you go. If there's something in the way, like a rock, tree, or the roof of a building, just say "Enhance it again" and you'll get all the resolution you need.
If that isn't good enough for you then maybe you could create a GUI interface using Visual Basic to do the job for you.
Would TV lie to me about this kind of thing?
Will old people at the bus stop be killed by predator drones because their walk is 95% similar to OBL's?
My own was using lamberts cosine law to gather angular information on leg position by the light patterns reflected off the thigh of someone walking directly towards the camera.
The problem with gait recognition is, AFIAK, it's not really been proven to be a decent biometric - i.e. I'm not sure it's really all that unique, not without measuring things at a very high resolution, which probably isn't going to be possible either from space or with the current install-base of cctv cams.
Anyway, scary stuff if it does work.
taste and smell..
"Excuse me,natalie portman, you have been selected for a random terrorist check, step behind the screen while our terrorism expert tastes you. "
"Does she taste like hot grits?! "
Read the first chapter or two of Cory Doctorow's Little Brother for some low-tech ideas on defeating gait analysis.
I like how yet again ``spotting terrorists'' is an euphemism for ``spotting everyone else, too''. I habitually substitute ``spotting YOU'' and honestly think of all the good it would do.
Yes, there's useful stuff in there, but again only if those watching you can be trusted. This has been said often enough before and still people score cheap headlines with the same fallacy.
Anybody spot the shadow of a flying pig yet?
So... um, you've got a "terrorist" under tight enough surveillance that you can build a "gait profile", but you're not arresting or just outright executing them?
Admittedly, I support this effort. Once complete, the DHS can take its rightful place as the Ministry of Funny Names and Walks.
c.
Log in or piss off.
So if it looks like a terrorist, talks like a terrorist, and now walks like a terrorist, it is a terrorist? We should find a way to see how they taste and smell..
I'm guessing - almonds.
If this were really happening, what would you think?
Yes, but that still doesn't answer his objections.
Let's say you've got nothing to hide, are on the database, then you get an abcess in your foot. I had one, for example, thanks to some retarded shoes which did that much damage and it got infected. Next thing I knew, my walking style could belong in a "ministry of funny walks" sketch, except for me it was more painful than funny.
Would I suddenly be outside the database, and thus a suspect, in that scenario? Or what if they entered a criminal in the database when he had a similar injury, and then I have a similar injury two years later and suddenly I look like the re-appearance of Abdallah ibn Jihad, wanted for arson, genocide and jay-walking in East Bumfuckistan and Elbonia? (Made up name, btw. Means IIRC slave of Allah, son of jihad, or enough to get your average anti-terrrist spook get his panties in a knot by itself.)
It's not like you can choose when you'll have such an injury.
What is the degree of confidence in such an identification, anyway? How fine you can slice a gait and still leave room for normal daily variations? (E.g., account for stuff like today I'm feeling chippy and walk a lot livelier, while yesterday was a shitty day and my walk probably reflected that. E.g., today I walk on grass in the woods, yesterday I was walking on wet concrete, and a month ago I was walking on sand at the beach.) As they say, "if you're one in a million, there are 6000 exactly like you." Will it be able to positively identify said Abdallah ibn Jihad, even when he's walking uphill through the snow with a pebble in his boot, or will it be more like "it's one of 6000 people, one of which is Abdallah ibn Jihad"? Again, that's the number if it could positively and unerringly distinguish between one million different gaits.
A polar bear is a cartesian bear after a coordinate transform.
Bullshit and strawmen, whether intentional or not. The objections to false positives have more to do with statistics, than with slippery slopes or anything else.
E.g., let's say I have a system which can look at photos from the security cameras, and tell you if a face or gait in the crowd matches a terrorist profile with 99% accuracy. (Which is actually a lot higher than what most of these snake oil systems get.) The problem in that case isn't that it lets 1% of the terrorists go. It's that it also creates 1% false positives out of people who aren't, for just one terrorist's photo. Apply that to just one airport, say the JFK, with its almost 60,000 passengers per day. If you get exactly one photo of each passenger, that's 600 false positives per day, in just one airport, for just one terrorist. But more likely you'll have everyone caught by several cameras during their trip to the airport, so the number multiplies accordingly.
Now feed it a database of several tens of thousands of known criminals, suspects, etc, and watch the number of false positives explode. Given that accuracy, just 100 photos are enough to match a majority of the passengers at one point or another.
At some point you can simply swamp the security with false positives, to the point where it's worse than useless.
And it's not just a hypothetical scenario, it's what airport security people themselves have said about previous trials with face recognition system. That they're crap and worse than useless. Would you accuse those too of being paranoid and slippery-slope types, or just accept that they probably know their job enough to know when a gizmo isn't helping it?
So basically spare me the bullshit about "nirvana fallacies" and "paranoid liberals". Learn what the real problem is, before talking out the arse about.
A polar bear is a cartesian bear after a coordinate transform.
I've worked on gait-analysis, along with facial recognition and other computer vision techniques. Gait-analysis is done by training an algorithm to identify a person's gait using a large amount of video as training data. You can't just snap a single picture of a terrorist and recognize their gate, or train an algorithm using 10 seconds of video that you have. You have to have sufficient training data if you want any meaningful recognition rate. As it is, gait-recognition has a much lower recognition rate than other vision techniques.
Making the training data useful for recognition is challenging enough. If you have footage of a person walking against a white wall at a controlled distance, it is easy to gather this data. However, if your training data is from a video of a person walking through city streets, much less a market place, there is an awful lot of human processing that needs to be done in order for the data to be useful for training. Also, as with facial recognition and other visual recognition techniques, gait-recognition is highly susceptible to changes in camera angle. If you train a gait-recognition algorithm on images of someone walking towards the camera, that doesn't mean that you can identify them with any reasonable success from the side or above. In essence, in order for this to work, you would need an ample amount of training data on a terrorist in a controlled environment. That probably isn't very likely. As you'll notice in from the article, these experiments were conducted in a specific controlled environment.
This story strikes me as someone doing some interesting research, but I'd be curious if we get any meaningful results from this work, even 10 years down the road.
Well, IIACVS (a computer vision scientist), so here goes...
No, this won't spot terrorists. Currently biometrics are pretty good at answering the question "is this person Mr. A?" where A is know, and providing a yes/no answer.
They're OK at answering the question "Which of A,B,C,D is this person?", up to a fair number of people.
What they suck at is "I A any of the potentially 6 billion people who might go past this camera?"
The reason they're good at the first, is becaus eif you want to get entry based on biometrics, you don't generally hide your appearance. In the last case, hiding appearance s easy. Basically if you wear a large sack and sunglasses and gloves, your face, irises, fingerprints and gait are not accessible. No fancy camerawork will help with that and the system will not work.
Then there's more minor things like beards/lack of for faces, and for gait, a stone in the shoe, leg injury, John Cleese, an embarrassingly placed itch, and so on which also throw off the system.
Basically, Humans have had millions of years to perfect (and a large chunk of brain dedicated to the task of) identifying people we know. We're eally good at it, and can identify people we know very easily in a reasonable sized group.
We still don't scale well up to very lage groups, probably because the problem is too ill poosed to be tractable.
SJW n. One who posts facts.
With homage to "The Usual Suspects," all the posters discussing ways to change their gait when committing a crime have it backwards. Start out from the first time your "identity" comes into existence with a limp, and after the crime is done, go back to walking normally. The gait marked with the criminal no longer exists.
https://app.box.com/WitthoftResume Code: https://github.com/cellocgw
Watching 6-7 billion people walk is out of the question for the satellites to cover. So, they specialize in spotting certain gaits before zeroing in and analyzing. Specifically, long, low strides, with one arm out in front, bent at the elbow, sashing a cape (possibly hiding a round, long-fused bomb). The other hand, if twirling a long moustache or rubbing the front brim of a black fedora, will tip off the satellite that it is, in fact, looking at a villain. The tracking of shifty eyes and maniacal cackling were removed for technological shortcomings... and the satellites kept targeting congress.
During testing, the engineers were proud to report the satellite alarmed them to several instances of women being tied to railroad tracks, banks being robbed, and suckers being stolen from infants. When a satellite makes a positive match to one of these terrorists, it will broadcast staccato piano music in a minor key to the area. Citizens are expected to boo and hiss these men if the satellites begin alerting them of their terrorist ways.
I am the richest astronaut ever to win the superbowl.
Your gait isn't so much like a fingerprint.
A little background. I am not familiar with this research specifically, but I have exposure to a very similar concept.
I used to do ground search and rescue in New Jersey. A big portion of what we did was woodland SAR for missing persons. Because of this, there was a lot of emphasis put on tracking.
There is a man named Tom Brown Jr. who is basically considered the modern foremost expert on the subject. He learned the basics from a young age from an old Apache scout who was his best friend's grandfather. It sounds incredible, but the man has written several books, both technical, and biographical. The technical ones aren't of much interest to folks who don't have an interest in tracking, but the biographical ones I would highly recommend. He currently runs a school in Toms River, NJ.
The organization I did SAR with put a lot of stock on Tom Brown's methods and incorporated them into their training schools. Eventually we opened up a dedicated tracker school, though I never participated in that level.
There is a technique known as pressure release tracking, where one looks at the characteristics of a track in a soft medium like sand, mud, or to a lesser extent, gravel or such. Within the track exists a whole environment that was created by the state of the organism that made it. Most people can figure out that if you shift your weight to your left, or favor your right foot, or are limping, that you'll see that in a track. But you can also see other things. Is the subject hurt? Is it hurt somewhere other than the legs? Is it tired? Is it male or female? Is it pregnant? How much does it weight? How tall is it? Is it carrying something? Does it have to urinate? Is it sexually aroused?
I know people who have reached the level where they can infer these things accurately. To me, it's not a stretch to believe that there are other ways that this could be done (this shadow technique for instance).
A good tracker can tell a lot by looking at your tracks, so I'd so I don't know how they plan to use gait data in a useful way, but I'm willing to entertain the idea.
Touch everywhere, even when inappropriate.