Re: Mysterious Faces

From: Dye, Matt (
Date: Thu Mar 20 1997 - 13:34:54 GMT

> In reference to the recogntion of faces, how is it that although we
> can recognise someone almost instantly when we see their face (assuming
> of course that we know the person) it is often very difficult to get a
> clear picture of the person in your mind (again assuming that everyone
> out there has a mind - I know I do) when they are out of sight?

It is also true that giving a description to the Police is difficult.
Police artists usually use a system of succesive approximation ... the
first sketch is normally way off target. Why might this be so?

If we used structural descriptions to recognise faces, would we have
access to a representation that could be used to give an accurate
description of someone's face? Yes?

If we used a "template-orientated approach" would it be as easy? No?

>Also, it sometimes happens that you see recognise someone but
> cannot put a name to their face. Is this perhaps a breakdown in
> communication between different modules of the brain?

Yes! It has been proposed that different modules deal with different
aspects of the problem. One module will tell us whether the face is
familiar or not, another will let us access the name associated with
the face, and yet another may allow us to access semantic information
about the individual. This allows us to explain deficits were someone
showed a picture of Margaret Thatcher may respond: "I know her. It's a
famous woman ... someone invloved in politics. I know I don't like
her. She was Prime Minister in the 80's. Is it Greta Garbo?"

If you're interested in this area, the names to look for include:

Vicki Bruce
Mike Burton

And keywords would be:

face recognition
interactive activation
covert recognition

It's all quite recent stuff, and not too hard to get your head around!

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:52 GMT