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."
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
This is old news - software that finds faces has been available for years. To cite an older example, the company Miros, which later became TrueFace - they used a neural-net approach.
I dont get the point ? It doesnt recognise faces, just tells you if theres one there. Thats not exactly state of the art is it ? When other companies are producing systems that can identify people from images, albeit inaccurately.
...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.
Almost all human faces have common characteristics, such as two eyes and one mouth.
The "almost" reminds me of the joke: "Do you know that you have more than the average number of legs?
Some people have lost one or both legs, but no one has three or more. So the average number of legs is slightly less than two."
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.
Steven Pinker talks about this in his book, How the Mind Works (New York: W.W. Norton & Co., 1997). He writes:
He also discusses a patient, LH, who was unable to recognize faces following a severe head injury, although he was in other ways entirely normal.It's important to note that this is a different question than the one the software addresses: it tries to distinguish which images are faces and which are not, not whose faces they are.
Seriously, that's old. I'm a computer vision Ph.D. student, and there now are much faster methods. I'll just refer to my old comment.
A demo can be found here. You can contact me for more details...
Current really fast methods use cascades of very simple classifier that are very weak themselves, but very strong when combined. The work of Viola & Jones is what most of the stuff is centered around nowadays.
Do your own here:
http://argus.cs.unimaas.nl/fddemo
Support a Europe-related section on Slashdot!
The human brain is not a general-purpose computer. There are specialised modules, which evolved by natural selection, for various tasks which were evolutionarily adaptive in the ancestral environment. Which is to say, if being able to perform one type of mental processing quickly helped your hunter-gatherer ancestors survive, find fit mates and not get cheated or otherwise bested by competitors, it gradually evolved into an optimised piece of neural hardware, its template coded in their DNA. Being able to recognise faces quickly (and thus be able to match an image of someone to what you know about them and their reputation) was a major advantage in a highly social environment, and so evolved into a highly optimised module.
Sometimes, through various disorders, these modules don't work properly. Which is why conditions such as autism (dysfunction of the relating-to-other-people module), schizophrenia (inability to distinguish between internal and external stimuli), face blindness (the face-recognition module). and so on, can exist. Sure, a face-blind person can step themselves through a face-recognition algorithm, but it's slow and laborious, and by the time you're done, that cute girl/guy you're wondering whether you recognised has moved on.