Slashdot Mirror
The Face Detector
Roland Piquepaille writes "Almost all human faces have common characteristics, such as two eyes and one mouth. Still, some people, affected by face blindness, cannot recognize one face from another one. So it's understandable that face recognition is a major challenge for computer vision systems. In "Facing facts in computer recognition,", the Pittsburgh Post-Gazette reports that a team from Carnegie Mellon University's Robotics Institute has developed a very accurate software to find faces within images. By analyzing only 768 pixels, the system can detect 93 percent of the faces in a set of images while falsely identifying four objects as faces. The Face Detector Demo is available online and you can submit an image for analysis and receive the results by e-mail. The technology will be used for security purposes, but also by digital photography companies who want to automatically reduce "red eye" effects. You'll find more details and references in this overview."
Where's Waldo? :D
instead of the endless "let's use it in airports" crap, it looks like they've found a real use for this kind of thing.
give it to blind people so they can know who they're talking to. But don't stop there - man, the number of times I've forgotten names... it'd be great if they could integrate this kind of thing into some glasses, that popped up the name of the person as you looked at them (assuming, of course, you knew them).
whoever commercialises that tech first is going to make a lot of $$$, I think...
-- james
...while falsely identifying four objects as faces.
That's interesting. The AI is sufficient to identify most faces. And it sees a few faces where none exists - not unlike people. Little kids point out when their bananas, carrots and peas line up just right to make a face. If they see it, why shouldn't a computer? What about the moon? Would this software see a face there? A man maybe?
There is no point to this, just interesting thoughts that struck me while reading.
Slashdot Syndrome: the sudden, extreme urge to correct someone in order to validate one's self.
Man, that would rock!
Though it might be a bit disconcerting to the people who observe you screaming "JACKPOT!" and jumping up and down on a seemingly innocent pedestrian.
Though face detection is easy for most people, some suffer a perplexing disorder called face blindness, or prosopagnosia, which is an inability to discern the differences between faces...One such sufferer, who is part of a research study by Behrmann, can't recognize his own children when he picks them up from day care. He relies on the day care workers to bring his children to him; failing that, he carries a "cheat sheet" of photographs that can help him make out who's who.
We just found a candidate for the Most Embarassing Dad of the Year Award!
Dad: Hi, I'm here to pick up my son, Billy.
Day Care Worker: Sure, which one is he?
Dad: Uhhh... (pulls out photograph) I think he's the one one the left... no wait, in the middle. I'm not really sure.
Day Care Worker: Uh, OK, sir, whatever you say. Let me just leave the room and, uh, get Billy. (leaves room, dials 9-1-1. A few minutes pass)
Kid at Daycare: Hey Billy! The cops are arresting your dad again!
The algorithm only recognizes where the face is. It does NOT recognize the face to match it with another picture.s ers/2236.html
The algorithm is almost scary, watch this sample
http://vasc.ri.cmu.edu/demos/faceindex/05062004/u
The problem is that even if you can recognize where this 80 pixel face is, it will be very hard to match it up against features of known people. Several [automated] face recognition systems implemented in Florida failed. In more than 3? months they failed to identify a single known offender.
Bruce Schneier points out the problem of false positives in his "Secrets and Lies", and you'll get the same argument in any freshman statistics class: If the target population (of identified terrorists in the country, or people with AIDS) is extremely small, the probability of a false positive is greater than the probability of a true positive. If this system is correct 93% of the time, it's wrong 7% of the time. How many terrorists are there in the average metropolitan airport? I'd say zero (on average) How many people would be incorrectly identified? 7%. An airport with, say, 2000 people in it would have 140 of them misidentified. Even if the average airport had 20 terrorists, the false positives or misidentified would greatly outnumber them. And since the system is wrong 7% of the time, one or 2 of those 20 terrorists would be misidentified as not being a terrorist.
I am no expert in this technology, but I am somewhat knowledgable about it, let me explain something.
You won't understand how hard is it to actually pull off something like face recognition until you yourself actually sit down and try it, only to realize that the problem is much more complex to solve when it has to be so all-encompasing.
The first step to face recognition is to recognize where the face is. The result of this process are quadrilaterals that carve out the face so that when you crop, you are left with exactly the face (frontal, or profile view or other).
A common technique used to do that is to locate the eyes. Most faces (heck, even those with veils on them for relegious reasons!) will contain eyes. Then, when detecting where the face is, you are only left with not having covered people who are wearing sunglasses (which are much easier to detect).
After you have located the eyes, you gauge by their proportions the approximate proportions of the face. Then, you apply an iterative technique (varies in principle, typically based on differential calculus combined with numerical methods of approximation) to locate the bounds of the face so you can eventually crop it to know WHERE THE FACE IS.
"Obviously", the iterative technique has to be able to detect false positives via a threshold set that will rule out the non-face. However, once you have located the eyes with certain reliability, the overall chance that you have come across a face is pretty solid.
The problem is complicated as it is already as you can see!!
Only after FINDING the face, you can start MATCHING the face. At that point you are facing a number of problems that the imagination of most /.-ers can conceive of... Bierds, smiles, teeth-showing, frowns, skin tone changes and the most popular by all scientist: plastic surgery....
A common approach to the actual face matching is a technique of the so-called eigenfaces, whereby you compute a "common" face of the pool and then you can navigate down the specialization of characteristics (e.g. bigger, bigger, bigger nostrils) as you drill down, narrowing down the pool of possible faces.
There is nothing that takes away from how much state-of-the-art CMU's research is. It would be like saying "why is someone dealing with virtual memory management of an operating system if by now, we already have user applications for the OS". Do you see the flaw in such thinking?
The science behind is a lot of mathematics, so dear parent, please don't be ignorant of this type of work just because you don't understand its complexities...