How do you describe a face? Given my ability to forget almost everybody I meet, it’s a question that bothers me on an almost daily basis. I’m always trying to figure out some procedure by which I can break a face down into its component parts, remember them methodically, and so be able to recognize somebody the next time I see them.
Oddly, I’ve not yet been able to find any systematic method for doing so. There are tantalizing hints that such systems exist, but they’re never spelt out simply on the internet.
In Snow Crash, Stephenson imagines the value of a reliable synthetic face for living in a virtual world:
He was working on bodies, she was working on faces. She was the face department, because nobody thought that faces were all that important – they were just flesh-toned busts on top of the avatars. She was just in the process of proving them all desperately wrong.
The Black Sun really took off. And once they got done counting their money, marketing the spinoffs, soaking up the adulation of others in the hacker community, they all came to the realization that what made this place a success was not the collision-avoidance algorithms or the bouncer daemons or any of that other stuff. It was Juanita’s faces. Just ask the businessmen in the Nipponese Quadrant. They come here to talk turkey with suits from around the world, and they consider it just as good as a face-to-face. They more or less ignore what is being saida lot gets lost in translation, after all. They pay attention to the facial expressions and body language of the people they are talking to. And that’s how they know what’s going on inside a person’s head – by condensing fact from the vapor of nuance.
That’s the dream, then. As for the reality: there’s probably something of that ilk in Second Life, but I’ve not yet hunted it down. The real action on the digital side is in computerised face recognition, which is alas of little use for people wanting themselves to describe faces. Early work was based, like old-fashioned anthropometry, on measuring the distance between ‘anchor points’ found in all faces. But as it’s moved towards more statistical methods, which get results but can’t be imitated by humans.
Meanwhile the police have procedures to help witnesses identify the characteristics of a face:
most composites are put together by asking a witness to describe the parts of a face — the eyes, nose, mouth or chin — and then assembling those pieces to create a likeness.
The popular FACES computer composite system, for instance, offers witnesses 63 head shapes, 361 types of hair, 514 eyes, 593 noses and 561 lips to choose from.
What are these part of the face? This paper contains a list, extracted from the Farkas System of facial recognition. The full list is apparently present in full only in Leslie Farkas’ textboook Anthropometry of the head and face
The Visage Project seems to be an attempted online classification of faces through identifying features. The demonstration is hampered by too-small images, but the descriptions of characteristics are useful.
Another branch of work looks at the face in terms of emotions. This area was spearheaded by Paul Ekman, an anthropologist trying (with some success) to demonstrate the similarity of emotions across human cultures. His Facial Action Coding System is a means of describing facial emotion, muscle by muscle. He’s spent the past decade training police, writing popular books, and even inspiring a TV series — but nonettheless seems to be a serious and broadly respected academic. The effort required to use FACS, though, is considerable — and it’s concerned with changes in exprerssion, not with the permanent structure of the face.
Finally, there’s a certain degree of scepticism about the idea of learning faces section by section. It isn’t normal, you see:
Several brain studies have shown that we tend to see a face as a whole, and we pay more attention to the relationship among the parts of a face than we do to the parts themselves.
“Every cognitive scientist who has studied faces has concluded that faces are processed holistically. In fact, we now know that at least as early as six months of age, babies are engaged in the holistic processing of faces, not individual features,” Dr. Wells said.
“You can take people who’ve been married 15 or 20 years and the husband or wife can be quite incapable of describing a single feature of the spouse’s face accurately,” added Christopher Solomon, technical director for a British composite company called VisionMetric Ltd.
Still, for now I’ll take the facial-component approach over the alternative of recognizing friends by their hair and shoes, and becoming confused whenever anybody has a haircut.