Slashdot Mirror
Face Recognition - Real or Science Fiction?
An anonymous reader writes "Facial recognition software has been touted as one of the technologies that will change our future, particularly in law enforcement. How close are we to being recognized by a computer anywhere we go, as portrayed in movies like Minority Report? According to the industry's recent Public Relations releases, these products are closer than we think.
The reality though, is that current products work only when utilizing a small comparative sample, and any attempts for an individual to disguise themselves typically throw off the results. To see how far this technology needs to go before becoming mainstream, one site utilized Government-tested face recognition software, available freely through MyHeritage.com, to compare hundreds of famous people, animals, and cartoons to a database of 2,000 celebrities. Some of the results showed promise for the technology, but most were just funny — for example, who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?"
"who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?"
I think it's more a question of 'how many beers' than of 'who.'
You mean Barbara Streisand and Shrek *arent* the same person!?
After working in computer vision for 5 years I've realized that most problems aren't hard - they are not well defined. Mathematically face recognition is not a problem that can be stated.
Many other problems in CV are like this - edge detection, segmentation, etc. But people write hacks that work in restricted conditions and say they've solved.
And look, you could always just put on those Groucho Marx glasses.
Surely trying to match people with cartoons or puppets is never going to work - what do they really expect. They sort of facial clues you get from a real person and very different to what you could get from a cartoon.
But, this does scare me - I invisage a future wear the government knows where you are at any time, if not by picking up your face on the streets, to embedding some sort of chip. This is the way we are heading (sure, not for a while, but that doesn't mean we shouldn't think 20 year ahead)
Just because its funny that Lance Bass has something in common with a Teletubby and Barbara with Shrek doesn't make it bad technology!
This is all well and good, but the minute I get falsely identfied as a criminal just for being in the bar district late at night in the wrong place/wrong time I won't be too happy. . .
disclaimer: I've been known to store numbers in my ass for which to dig out when quantities are required.
So I guess next time a teletubby or Shrek wanders through a mall, they're totally going to throw off the face-recognition software.
Is it just me, or does that seem like a stupid way to test the software? If you want to show that rudimentary disguise is an easy way to get around it, that's valid, but just messing with the sample of potential matches by throwing in cartoon characters destroys the validity of the "study".
-stormin
The Southern Baptist Convention has creationism. On Slashdot, we have porn.
>> who would mistake Barbara Streisand for Shrek
Actors hate it when fans think they are the characters they play.
I thought they used chips in the eyes of people in minority report, not face recognition.
No kidding on the oops part. Taco, you put the link to MyHeratage as a relative URL instead of an absolute one.
0*0
00*
***
who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?
Time for an "iwould" tag.
I've tried out the software and it was fun for some laughs. I'm not sure how it works exactly but I can tell that the angle of the face makes a difference. When I put one picture of myself in where I'm looking ever so slightly to the right, I'm matched with celebrities photos looking in that direction. When I put in a similar photo facing the other direction, I get a different set of celebrities looking in the other direction. There's a few overlaps and those are the ones I think I look the most like (although it's a stretch to say I have anything that could pass as a celebrity look).
Not to nitpick excessively, but you could easily substitute portions of this article with terms like (and relating to) “Internet”, “personal computer”, “telephone”, “car”, and others. Asking ourselves if a technology is “real or science fiction” when it already exists (albiet in a primitive form) is silly. Of course it exists; the question itself cites examples. Perhaps the meaningful questions might be along the lines of: “what are the challenges associated with making it accurate?” or “what impact will facial recognition have on society?”
Why bother.
I'm wondering about the legality of all this, especially in a criminal justice system. My DNA, for example, can't be used in court as evidence unless certain hoops have been jumped through; the prosecutor needs a reason to obtain a DNA sample and then procedures must be followed.
I wonder if the same systems will apply to a computer analysed image of my face; will there be a criterea for when this image is admissable in court? Will I have rights concerning my image? Or are we just going towards a 1984 style system. Interesting because this hasn't been the result of DNA admissions to court, despite the seemingly more robust nature of this evidence.
For example, who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?
So, i see it's working correctly!
Humans seem to have no problem with it, so it's clearly possible.
Bill Gates seems to look like... Meryl Streep???
Nuns. No sense of humor. -Kurgan
Isn't this the sort of problem that might be well-addressed with genetic algorithms? We've got a problem that we can't really define well (mathematically), but we know what good results should look like. So the actualy process to solve the problem is hard to design, but a test for good results is (comparatively) easy to design.
That sounds exactly like the sort of problem that you could use GAs on.
Unless, of course, I'm completely wrong about the state of the art in genetic algorithms, or am making some other fundamental error in my reasoning.
Reality has a conservative bias: it conserves mass, energy, momentum...
Evidently, I look a lot like Tom Cruise. Well, actually, the first hit was Matt Stone, but with a much less flattering picture. Tom Cruise was 4th or 5th on the list, after Luke Wilson. Here was the picture I submitted. Naturally, I'm being sarcastic when I say it's flawless.
Ben Hocking
Need a professional organizer?
Okay, so I look like your regular geek. I'm fat and bald. Yet I apparently look like some hot female movie stars. 80% like Grace Kelly. Not sure if I can link directly to a result. Let's try: http://www.myheritage.com/FP/photo.php?siteID=1&ph otoID=5969307&source=album&sourceID=963790&albumID =963790
Granted - the examples looked pretty good, but I just can't see Grace Kelly when I look myself in the mirror.
I believe Minority Report used retina scans, but that nit aside facial recognition works to a degree and will only get better. Security cams will eventually upgrade to HDTV resolutions, perhaps augmented with very high resolution stills when a potential match is made. This will all take more processing power, but all mighty god Moore will eventually gives us this day our daily CPU load.
About false positives. So what? Eyewitnesses make mistakes also. Eventually, perhaps very soon, machines will surpass humans in this arena just as they have in others. Can anyone here on Slashdot defeat Deep Blue at Chess?
As to the legality or ethics, what can be done will be done, at least in public areas. If it would be legal for a human to do (they haven't outlawed humans scanning for suspects in public areas) then it will be legal for machines to do despite the unease many will feel knowing they are constantly being watched.
Letter To Iran
Hmm.. Well Myheritage seems to think I look like Chevy Chase and John Travolta.
When in fact I KNOW I like like Prince William.
Sounds good. You could use GA guided neural networks. Worked once (i.e., us), so it stands to reason it could work again. I'd set up an input system similar to how we understand V1 to work, and then let the GA explore over a range of options for V2, V3, and V4. Each of these would be constrained by some of our best guesses of how they work in humans. I've been trying similar things recently with a model of the hippocampus.
Ben Hocking
Need a professional organizer?
Now we have those banners advertising magical spraypaint-thing that makes your license plate harder to see through a traffic camera. I guess in the future, we'll be seeing banners advertising stuff that you spray on your face so that nobody recognizes you.
This is a sig. It is appended to the end of comments I post.
Here are two pictures of Matt Stone, one of the pictures of South Park, just to prove your point.
Ben Hocking
Need a professional organizer?
The facial recognition on myheritage.com is a complete joke. All it does, for the most part, is match skin tone and face angle in relation to the camera. I hope to god that whatever else out there being developed is a little better thought out, you know, something actually based on some sort of algorythm that wasn't designed by a third grader.
But don't we almost always get a computer to solve a problem that's not strictly a mathematical one using "hacks that only work in restricted conditions"?
Our spell-checkers in our word processors don't actually know anything about the rules of a language, phonics, etc. They just do lookups from a dictionary. If a word's not listed, it has no idea if it's spelled properly or not -- even if the misspelling is one that's simply not a possible correct sequence of letters for the language. Most don't even realize if a word is misspelled in the context of the sentence, as long as it matches a correct spelling in the word list.
Until we figure out how the human brain recognizes faces as individuals, we can't expect anything *but* a clever hack for a computer to do the same. And truthfully, I suspect the human brain takes many things into account to do a "recognition" on a person. How often do you see somebody in the store that you're pretty sure you know from a previous job, school, etc. but you're not quite sure? I've had this happen a few times, and to make a better determination, I had to take other factors into account, like the sound of their voice if I heard them speak, the way they walked, or maybe an expression that came across their face. Humans "key in" on specific things that help them remember a person. And depending on which "features" they chose, they may or may not be effective. (Say you remember a gal really well because of her long, flowing hair? If she cuts it real short, there's a good chance you won't recognize her at all anymore if she walks by you.)
who would mistake Barbara Streisand for Shrek
I do that all the time.
This technology has been used in Vegas to identify cheats for a long time. I guess the difference is its looking for particular people instead of trying to identify every person. More at http://www.cnn.com/2006/TECH/07/13/popsci.gambling /index.html/
I tried the http://www.myheritage.com/ site to do the funny who you look like example. I used two face photos one (3.2 MP) with the ocean behind me and one in a park (6MP). It couldn't find a face in either photo. If this is the best this technology has to offer we are many years away from anything real. I know the US government is using this stuff, but they are still looking for WMDs in Iraq if my experience is typical with the same results.
My two cents,
verl
Cartman: Try this on for size. Blood-drenched, frozen tampon popsicle!
Sadaam Hussein: Hey buddy, I know I was mean before, but don't worry, I can change!
Cartman: Okay.
Not. Fuck, shit, cock, ass, dildo, boner, bitch, pussy, butthole, Barbara Streisand!
What about plastic surgery? Identical twins? Even a close sibling could be similar enough to fool software in some cases.
On the other hand, fingerprints are completely unique, even between identical twins, and (last I heard) unchangeable. Researchers would be better off spending time on improving fingerprint-scanning technology for identification purposes, although clearly face scanning would be even less intrusive for other tasks.
Every morning I wake up to look into the mirror and it's a different face that I don't recognized. Maybe I need to upgrade my mirror?
Showing the resemblance in absolute numbers, better with some P-value or E-value.
I do not believe in karma. "Funny"=-6. Do good and forbid evil. Yours, Oft-Offtopic Flamebaiting Troll.
To be fair to the software, I can see the resemblance!
However, as other posters have noted, the match seems to be more based on orientation of head, facial expression, etc. I don't see much facial recognition going on, just primitive image matching.
And how the hell you got modded off-topic is beyond me. Thanks for posting a link that let me see this garbage in action without having to mess around uploading my own pictures or jumping through any hoops.
I currently work for one of the larger digital biometrics company and our facial scanning software is quite impressive to see in action. The problems companies face right now is crowds of people. Facial scanning systems are best in areas such as airport checkpoints because it's an easy place to limit confusion with the algorithms. There's some very cool stuff we have in the works but I'm not sure how much I'm publically allowed to give information about.
I don't think facial scanning will ever be used for convictions but I don't see why it shouldn't be used for finding suspects. Imagine a bank is robbed and they take video. Digital biometrics analyzing against a database of known criminals is no different than having an employee look through a mugshot book. It's faster and then can be cross checked with eye witness accounts.
http://www.riya.com
Even the "demo" requires registration, including name, sex, and birthdate. That's just plain wrong.
to err is human, to forgive is divine, to forget is... umm...
If I was training to match V1-4, I'd have the input come from two "eyes" with inputs similar to what our eyes actually provide to our brain. We know quite a bit about visual cortex, but there's a lot we don't know. Initially, I'd train it using a batch of photographs for a single person (we'll call her "Momma") and then I'd train with a few others (where a match is a match only if it's the same person). From there, I'd create histograms of parameter settings that seem to do an adequate job on this small set, and then use this reduced parameter space to create populations that are evaluated after training on millions of photographs. (The photographs can be placed in front of the eyes - once for each photograph, mind you, and not for each "individual" being tested - just like we can recognize photos and not just people.)
I could imagine narrowing the parameter space down to 100 or so unknown parameters, and each training session might take several hours. Given enough resources (e.g., the Pittsburgh Supercomputer Center), I'd run population sizes of 500 or so (in parallel), so that you could possibly go through 4-5 generations per day. In a month, you might have some pretty good individuals. Of course, my research area is the hippocampus and not the visual cortex, so it might take significantly more than 100 parameters to even begin to set this up.
Now, someone else pointed out that such computers would not have the biases that we humans have, but that's not necessarily true. If you train the computer using an input set of 950,000 "white" people and 50,000 "black" people, it would tend to make the mistake of thinking that "black" people look a lot like each other. (Studies done with speech recognition have shown that neural networks trained on Japanese have a much harder time telling "l" from "r" than those trained on English.)
Ben Hocking
Need a professional organizer?
"who would mistake Barbara Streisand for Shrek, or Lance Bass of N'Sync for a Teletubby?"
Tell that to Cheney, who has trouble distinguishing between a man and a duck.
wear a mask! Flaws of this technique: 1-There are many realistic mask makers out there (should I say Hollywood?). 2-Twins (especially evil ones). 3-Changes in the face (shaving, burning, etc.). . . . I can't see this technique applied anytime soon.
Syllable 0.62 is here at last!!!
We simply need strong pattern matching algorithms for images...then anything could be recognizable. The brain uses pattern matching, doesn't it?
"How close are we to being recognized by a computer anywhere we go, as portrayed in movies like Minority Report?"
Now I could be wrong but I am pretty sure Minority report was portraying retinal scanning not facial recognition
But don't we almost always get a computer to solve a problem that's not strictly a mathematical one using "hacks that only work in restricted conditions"?
.
Out of the mouths of . . .
Our spell-checkers in our word processors don't actually know anything about the rules of a language, phonics, etc. They just do lookups from a dictionary. If a word's not listed, it has no idea if it's spelled properly or not . .
Because the standard spelling of a word is simply and precisely mathematically definable, whereas language is not. That is the parent poster's point. We call them computers for a reason. Deep in their little rocky brains all they can do is compute by shoving "beads" around.
Until we figure out how the human brain recognizes faces as individuals, we can't expect anything *but* a clever hack for a computer to do the same. And truthfully, I suspect the human brain takes many things into account to do a "recognition" on a person. How often do you see somebody in the store that you're pretty sure you know from a previous job, school, etc. but you're not quite sure?
In other words, face recognition doesn't actually work. Even the human brain is horribly inaccurate at recognizing faces. Anyone working in law enforcement in any capacity is well aware of this. That's why we now use fingerprints and DNA to more precisely define an individual . . . or electronically tag them.
I used to work in this field, but had the advantage of the full cooperation of the subject in desiring to be recognized and the only actual solution is to uniquely tag each one with something that can be read from a distance and looked up in a database. Assign each one a "word."
Why do you think racing cars have numbers painted on them?
KFG
The free software from myheritage does not compare to what is being done by several private companies associated with/funded by law enforcement or the military.
This is not a self-referential sig.
Facial recognition software has been touted as one of the technologies that will change our future, particularly in law enforcement.
... I can already aee it being implemented everywhere from everyone's home to the military and government entry security. ... And then MythBusters coming and defeating it with a pencil drawing of a face done by a retarded kid in a hurry, just like that episode with the fingerprint readers.
What a bunch of bull. Sure it will,
Think about "voice typewriters." Then think about what we really have. Yes, speaker-independent voice recognition systems that can recognize the words "yes" and "no" and the digits from 0 to 9 exist, and work reasonably well for short strings of digits like ID numbers that can be read back to the caller.
No, we do not have voice typewriters, and if you don't believe that, well all I can say is, "dear aunt, let's set so double the killer delete select all."
A face recognition system that could be mounted in a dashboard and recognize whether it's my wife or myself sitting in the car, and call up and power-adjust the seat and mirror positions, is perfectly feasible... it could probably be marketed today as a $10,000 car option if there were enough yuppies willing to pay that price. A toy. A convenience. And no serious downside if the system doesn't always work perfectly.
A face recognition system good enough for airport security or "global security concerns" is nonsense.
"How to Do Nothing," kids activities, back in print!
reading up on how cognitec (the folks supplying the facial recognition capabilities) do the facial recognition, i'm utterly unsurprised by these results.
in a nutshell, they apply face detection to localize faces in the image (good, although they're clearly ignoring the flesh color cue since that would have immediately ruled out shrek), then they scale and rotate the face to a fixed, standard position based on the position of the eyes (since they only have two points, they're absolutely limited to planar rotation and scaling, which means that views that aren't head-on are going to cause problems for recognition unless you have a really large training set covering a wide variety of poses). if you think about it, the scaling step probably explains the lance bass/teletubby and barbara streisand/shrek correspondences (squish or stretch the faces and suddenly they look a lot more alike). finally, they feed the corrected faces through the recognition engine which appears to be a k-nearest neighbor type classifier. this is rather odd since facial recognition has traditionally involved using techniques like eigenfaces and more structural / 3d approaches.
also, if you read the performance evaluation done by NIST and DARPA (available here http://www.frvt.org/FRVT2002/documents.htm and look at the ROC curves, it become pretty obvious that the performance of these algorithms is, in general, abysmal.
Answer: just put up a site where the public submit their own photos (after registration) to see which celeb they look like :-)
Steve
Any sufficiently advanced bug is indistinguishable from a feature.
Tokyo train station gets facial scan payment systemsa tion-gets-facial-scan-payment-systems/ [engadget.com]
u r_face_could_be_barcode.htm
s _kasumig.html
http://www.engadget.com/2006/04/27/tokyo-train-st
Your face could soon become just another 'bar code'
SURVEILLANCE CAMERAS AT STATIONS
http://www.infowars.com/articles/bb/biometrics_yo
Tokyo's Kasumigaseki Station
http://www.smartmobs.com/archive/2006/04/26/tokyo
Previously: "Linux... Toward the Sunrise..." Now: "Linux... Toward the-- No, now, part of Every Sunrise"
So not only do I not believe claims that facial recognition "is coming soon". I don't believe claims that it's "already here". I don't even believe the demos done right in front of my eyes, because we all know what demos are like. I'll believe it when a credible organisation trusts their money or security to it for an extended period.
Doesn't it make you feel good to know that our freedoms are protected by politicans, lawyers and journalists.
Gimp your eyes.. one up and one down. Round out your mouth like Nancy Crater (the Salt Monster from Star Trek's "Man Trap" episode). Be sure to put a stringy mop on your yead. Add a couple of black dots or raisins to your face to see if actors with moles come up.
Or, adjust your hair with hair glue. See if Sid Vicious or Suicidal Tendencies or the like appear...
Previously: "Linux... Toward the Sunrise..." Now: "Linux... Toward the-- No, now, part of Every Sunrise"
Teleportation is science fiction. Antigravity is science fiction. Time travel is science fiction.
Face recognition is definitely NOT science fiction. Humans can do it, as can chimps, dogs, dolphins, horses, to some degree. Computers can sort of do it in restricted situations. Clearly no laws of the universe need to be violated in order to make an effective computer face recognition system...
My bicyles
It doesn't matter how good the automatic recognition is, the false positive problem will always dominate the performance. Suppose it is 100% accurate at recognising bad guys, but generates a false positive 1 out 10,000 faces? In a busy airport that may mean 20 or 30 false recognitions a day and each may require a full security alert.
Near Tampa put a system into practice a couple of years ago. First time out of the barrel it IDed a "criminal" who turned out to be someone else.
Sorry, no sale.
Did it mix up Ted Stevens with Grandpa Simpson ?
until (succeed) try { again(); }
Chalk up one more reason to wear a headscarf/burka/etc., even if I am a guy:-)
This is all probbly moot at this point
Now, you can get a face transplant and be whoever you want to be anyways
LONDON (Reuters) - Surgeons were given the final go-ahead on Wednesday to perform the world's first full transplant, a radical procedure that has raised concerns about its physical and psychological risks. The UK Face Transplantation team at the Royal Free Hospital in London received permission for four transplants from the hospital's Research Ethics Committee.
use retinal scans for all of the identification needs? Similar, but != to face recognition.
``Our spell-checkers in our word processors don't actually know anything about the rules of a language, phonics, etc. They just do lookups from a dictionary. If a word's not listed, it has no idea if it's spelled properly or not''
That goes for spell checking in general, also when it's not being performed by computers. Whether or not a word is spelled correctly is exactly equivalent to whether or not that word exists in a dictionary containing all words in the language...and no such dictionary exists. The best you can do is declare some dictionary to be the authorative source at some point in time, and then you can judge whether or not a word exists in the language of that time. That's a task where computers probably beat humans.
Please correct me if I got my facts wrong.
A system like this doesnt necessarily need to be worried about false positives as much as it has to be worried about false negatives. While a 100% accurate system would be great, if they can guarantee that the system will never return false negatives, and will only return false positives 1 out ot 10,000 times, the system would be great to use not as a final judgement, but for leads. If every 'positive' were taken with a grain of salt and scrutinized by the authorities then it would be no different then the thousands of people responding to wanted posters. The only difference is this would be quicker and would give other information such as exactly where and when the sightings occured.
First post = troll. Cleverly worded post designed to enrage others = flamebait.
Under various circumstances, even perfectly functional humans have trouble interpreting what they see. Camouflage and optical illusions are just two examples that I can think of. Think of the drawing with some circles and some lines that asks, "which circle is larger" and it's clear what the answer is, but when you measure them, they're exactly the same size.
I don't really have a point, I'm just adding human cases to the statement that computer image interpretation is not well defined.
- Andy
Move all sig!
There are known solutions to this problem, we were looking at the "eigenfaces" method in class the other day.
There are some things though, first, the picture needs to be normalised: i.e. it needs to be set to the appropriate scale, and the head has to be at the correct angle (usually straight). This is why Bill Clinton with his head tilted to the left and looking to his left in one of the MyHeritage.com examples doesn't match the other picture of him with his head straight and looking towards his right; except at the fourth attempt (which is fairly impressive, BTW).
The next thing is that if you disguise your face with big glasses and beards, a computer will be no better able to do the job than will any human. They are DISGUISES after all, they hide detail.
The technology still has advantages though. In any UK airport, you're asked to stand in front of the customs desk and look at a camera: this takes a picture of you, and thanks to the instruction, it's to scale and your face is straight. I'm not sure if they do this at the moment, but the software could then look up a couple of thousand people in seconds using this technology, and submit likely possibilities to the guard who can then use his own eyesight and intuition to see if any of these are matches.
It's a whole lot better than training hundreds of guards to recognise thousands of faces.
Yeah, I looked at the gp post and Grace Kelly. The follow features stood out as being very similar: eyebrow line, mouth shape (and position), and the nose (although they pointed in different direction and were different length).
Layne
According to the book "Bringing down the house", facial recognition software has been in use in Vegas for 6-7 years to help identify known cheats and card counters. Software has likely advanced greatly since the book was written.
One interesting tidbit was that ears are often the best indicator of a match. They are not normally bothered with when people put on disguises.
Camera lenses will break and computers will bug when recognizing certain faces.
I'm afraid I'm going to call shennanigans on some of this. I've been doing Vision work for about 5 years now with a hefty does of image and signal processing in the mix(Working as gradstudent in the field right now in fact). Edge detection is well defined. The canny and shah-istan(think that's the name) are about as close to a mathematical optimal edge detector as one can get. There is in fact a well developed body of theory regarding differentiation of Signals. The problem doesn't lie in the mathematical models involved. It lies in how many people want to use those models. Edge detection suffers from spurious edges or edge flakes which are a symptom of noise in the signal at differention(ie differentiation enhances noise, integration smooths it). Segmentation can also be well defined you just have to be clear on what it is you're segmenting. Are you working in a color space, texture, motion? That matters. However you can get some very good results in these fields. See GPCA techniques for some examples of doing it. Or even modified PCA + EM or PCA+ Kmeans(clustering theory). Again very well defined. Mathematically there are several models for face recognition. One can examine the ideas of eigen faces(not my personal favorite but it's there), kernel based SSD type approaches to find key points, partial face detection followed by recognition over a sequence of images used to reconstruct the face, and more. The problem isn't the math. It's that when you project a model you are essentially destroying an entire degree of freedom which is a huge deal. Further just as you can match a partial finger print or a partial ear print you can match partial facechunks. The problem with makeup or facial hair comes when one relies on global matching techniques or uses only 2d information to do the matching. Now I'll be a first to say that alot of computer vision is a solution in search of a problem or that people do use a number of cheap hacks and dirty tricks to get things working but saying it's not mathematical is a lie. I can turn around and see at least 3 books at a glance that detail the mathematics that are a part of vision and image processing. So please don't confuse peoples fuzzy use or lack of understanding of the math for there being no math. Note: Machines are also bad at a number of tasks humans are really good at but the same can be said that there are many tasks that humans are very bad at but the machines excel at. Absolute range detection is a good example. Humans are very bad at telling you the exact range to an object, even with some sort of scale of the scene reference. Computers on the other hand(while suffering from noise in the signal) are still able to achieve significant accuracy depending on the range. You can see tyzx for an example of a comany who makes highly accurate stereo rigs.(They were around as of 2 years ago at least and I assume they're still going strong) Cheers
I don't care what you say, all I need is my Wumpabet soup.
I've seen and played around with a facial recognition system that could correctly identify emotions and would draw a box on screen around each face in the image (done in real time with a thermal camera, works even in the dark!) and then write the proper name for each emotion expressed under the box along with estimated distance, etc. It was very accurate, and fooling it, while possible, required a good deal of concentration to express the proper emotion, and without seeing the screen, chances are the emotion you were actually expressing would slip through at some point, so you couldn't completely mask what you were thinking. The question is, will you be faking the expression at the moment you are caught on camera? Will you be able to hold the fake expression (requiring a lot of concentration) and do anything else at the same time. Throw in an RFID id card system and they will be able to know who is where and what they are thinking, anytime you are on camera. So microwave those RFID tags and wear paper sacks over your heads! (Ski masks and panty hose are really suspicious and still could give away identifiable expressions!) Even worse, the real time emotion identification could be used to identify thought patterns with a decent accuracy (at least with the thought patterns you want to find and stamp out...)
Last Post!
Wow, maybe it does work...
A pic of G.W. got a match to Adolf... and a pic of hot grits matched to Natalie Portman.
You can have it fast, accurate, or pretty. Pick any 2.
As long as we get "probable cause" out of it, it's good enough.
"Reality is that which, when you stop believing in it, it doesn't go away." - Philip K. Dick
What we need is a gadget running this software that fits on a pendant, so people with prosopagnosia can figure out who just came up to them and hugged them. "Shrek! So nice to see you again!"
"From today [October 05, 2006] all British citizens renewing their passport will receive a new biometric passport. The hi-tech, or ePassports contain a secure chip with an image of the holder's face, and are designed to make forgery more difficult and improve international security ... also from today passport fees will rise. A 10-year adult passport will now cost £66" - Times
Q. So why are we paying this tax on holidays for a technology that apparantly couldn't tell Osama bin Laden from Captain Birdseye?
A. Although "facial recognition biometric data" all sounds very sci-fi, it is in fact the least accurate biometric identifier there is, according to experts. It will, however, be good enough for entry into the US without a visa - BBC
Reduce, reuse, cycle
Obviously facial recognition is a bit more complex than my example, but, the process is similar. Programmers take on a task, realize how much work it is, find a quicker way to go about doing it, and then take that shortcut that dooms them into claiming victory.
In my example, I wrote a program to play Solitaire. It used a primitive form of optical recognition to identify numbers 1-2-3-4-5-6-7-8 and rest of the board from examining screenshots. The algorithm was a bit slow (after all, this was only a prototype) but it did the job just perfectly.
Then I realized that I could take a shortcut. In Windows' Solitaire, the field numbers are color coded; therefore, I reduced my "pattern recognizing" algorithm into a color-recognizing one and improved gaming efficiency at expense of being cheap about it.
After I had developed rest of the game and made it work, I realized that couple of situations were unaccounted for : blank fields, and marked fields. At this point in time, it was miserable to go back and try to rewrite the recognition algorithm to help identify the board so instead I started taking more shortcuts and making assumptions.
End result? Lets say that a human is still reigning champion of the Solitaire world.
I haven't read the article, but I've used MyHeritage before. It doesn't do a very good job at all. You can tell it takes cues off incidental things like: which direction the light source in the pictures are coming from, what angle your face is at, how much are you smiling, and what style of glasses you're wearing. Good facial recognition software needs to be able to ignore these things and look instead at the actual shape of a face.
-- dR.fuZZo
For the most part, vision recognition technologies have acquired a circus type appeal; the technology is being manipulated for generating media blitz and courting VCs. Several startups utilize this type technology for focusing on color hues and various traits, but dodge the terrific challenge of recognizing images in its entirety. The interesting question is, with all this noise, to what extent will it slow down innovative technologies from coming forth.
I'd go for 3D skeleton recognition, if there's a way to scan them without large doses of x-rays.
It would be hard to fool a system like that. Skeletons usualy don't change a lot over time, and the database can determine how much change is acceptable, based on the time difference between old and current scans. As far as I know skeletons may be just as unique as DNA or fingerprints.
This is not the sig you're looking for.
Guardia (http://www.guardia.com/) is a Danish company that has been working on this for several years. They were featured in a show about startups last year. I think they use two cameras and infrared scanning to make the system more reliable, although I'm not sure how well it works yet.
Their floor model looks pretty neat. It was featured in the show. The cameras are mounted on two pillars and can move up and down by themselves. You get the feeling of a creepy intelligent robot, not far from Minority Report.
Anyway, they are working on it. If it can be done and they don't run out of money, I'm sure they will succed. The entrepreneur (the CEO) is really dedicated to the idea.
Yes, but that's not exaactly my point. I'm not arguing it's a very effective and efficient way for computers to verify the correct spelling of words. We've been doing it for at least what, 15-20 years now, in our software packages?
My point was more the fact that a human would usually catch certain errors that a spell checker won't necessarily catch. For example, in the English language, we know that q is always followed by u in our words. But a spell checker has no such information coded into it. It probably "just works anyway" in this situation, because we don't have that many words with "qu" in them, and a decent dictionary of word spellings will likely cover all of the possibilities. But it doesn't make it any less of a "hack", as far as how the computer solves the problem (brute force comparisons from a predefined list).
> Until we figure out how the human brain recognizes faces as individuals, we can't expect anything *but* a clever hack for a computer to do the same.
There are many ways to do long division. Computers don't need to mimic the way humans do it in order to be doing it correctly.
``My point was more the fact that a human would usually catch certain errors that a spell checker won't necessarily catch. For example, in the English language, we know that q is always followed by u in our words.''
I understood that, but I don't think it works. At the end of the day, spelling is _not_ bound by rules that are consistently followed in most languages. Even if it were in English, you would run into trouble with things like names and loanwords. For example, you say q is always followed by u, but it's not:
al-Qaeda
Qatar
QED
qty
qwerty
So you must, at least, have a list of words that are exceptions to the rule. You could, of course, use the rule to let the spellchecker give responses like "I don't know the word, but it looks ok" or "I don't know the word and it looks wrong". However, at the end of the day, whether or not a word is accepted into the language depends on consensus, and not (ultimately) on a set of rules.
Please correct me if I got my facts wrong.
When they can recognize my face through an Alfred E. Nixon facemask, I'll be impressed.
... or will *ever* be able to do ... where's the beef? I don't believe 90 percent of it. It'll create so many false positives that even a Machiavellian funding level won't provide the needed personnel to sort it all out.
What'll they do, switch to positrons?
As for the claims about what the technology can do
And of course, the "free" market will respond with stealth makeup. Call it "counter-physics".
"You must try to forget all you have learned. You must begin to dream." -- Sherwood Anderson
you recognize the camera (into SONY, NIKON etc...)
$150,000 for a face recognition system. It ran for two years, seeded with all known local offenders. It never found one despite the fact that cops caught some in the areas of cameras. They shut it off after two years of failure.
"I may be synthetic, but I'm not stupid." -- Bishop 341-B
I have Aibo robot dogs that can distinguish my own face and voice from other people, and they are essentially toys. It's not perfect; the camera is low resolution and the software can be fooled by glasses or hats... Even a hairstyle change can throw it off. I imagine a more complex system could make much more reliable determinations, and that is something that is likely to happen at some time.
Couple a good camera to a robust analytical program and give it access to the ID database and this is a probable future capability.
You don't need high tech for false positives.
One of my former boss has the same name as a criminal WHO IS STILL IN PRISON and it is a real PITA every time he has to take the plane.
I have prosopagnosia, you insensitive clod!
True. Facial recognition is a heck of a lot more difficult if just one of those "brain hacks" has failed to compile.