Lawrence Stark Interview: Conversations with History; Institute of International Studies, UC Berkeley

The Mind's Eye: Conversation with Lawrence Stark, M.D., Professor Emeritus of Engineering and Optometry; 8/16/00 by Harry Kreisler
Photo by Jane Scherr

Page 4 of 6

The Mind's Eye: Applications

In the context of your theory and what you've just described, what is it that a great artist does? A painter such as Bonnard; a photographer such as Gilles Peress. If they are trying to capture a scene or an image, what does their work do to us and how does the mind's eye of the viewer respond?

It all starts in the mind's eye of the artist, where the artist has a conception of what is the picture. What is a picture? See, the picture isn't just something you point a camera at. A picture done by a great photographer or an artist is somehow a selected, minimized picture. It has certain points of view, certain regions of interest that illustrate to the artist what he has in his conceptual mind's eye. So the first thing the artist has to do is have an idea of what he wants to make the picture of, what he wants to compose it of. The second thing is he has to be very, very technical, have great technical proficiency, to be able to paint what he wants -- the enigmatic smile of the Mona Lisa. And then that painting is there as a message, as a communication to all the other human beings who live at that time, who can see the painting, or who will ever live, as long as the painting is there. The person viewing it looks at that picture and has a mind's eye view of that and he can look at different parts of it and siphon up different regions of the picture in high resolution. But that's a very time-consuming and brain-consuming process. We make about three glimpses a second only, and those are very small areas of the fovea, foveation. So we generally pick up five to seven points of interest that check on what we guess is there. And then the mind's eye of the viewer may develop the same kind of emotional impact on his brain that the mind's eye of the artist expressed in his emotions.

What is great art, then? Is it great because of the efficiency and the economy and the universality of what it captures?

I'm not sure of the answer. You've given what I think is a good answer in your question. But it must [have] some efficient communication, and in a language that people can understand. In a language that's visual, yet motivational; emotional at both ends.

How do emotions enter into this process you're describing? In other words, is it purely efficient, pragmatic, or what do we know at this point about emotions and their relation to the physiology of this mechanism?

I don't think much is really known about it. At least, I don't know much about it.

You had some interesting things to say about virtual reality. You asked yourself in one of your works, "Why does virtual reality work? Because reality is virtual." I assume that comes from your understanding derived from the scan path theory. What did you mean by that?

It's actually from a different series of experiments. People have wonderful, computerized displays which we're all familiar with now. I mean, it started with storytelling around a fireplace, and then books, which let people's imaginations very free, and then with a plays and movies, television, and now we have what we call "head-mounted displays."

Virtual reality is essentially a couple of TV screens in front of the eyes that are arranged to allow three-dimensional viewing. These artificial worlds that are put there are very compelling for several reasons. One is that if you move your head, they can measure the movement of the head and they can change the picture. So you feel like you're moving around or looking around in a real room. If I had a three-dimensional model of this room, with you and me and the cameras in it, and I had a helmet-mounted display, I could look around and it would give me that three-dimensional picture, and it would also give it to me as I moved around. The human activity gives it a validity. And also if we measure, say, where the hand is, and I move my hand up in front of my eyes and an artificial hand moves up in front of my eyes in this artificial three-dimensional world, I start to believe that I'm in this artificial three-dimensional world. That's what they call "virtual reality," and it's used widely, not only in entertainment, but in the workplace now. A number of my colleagues have worked in that area, as I have.

The thing that makes me think from an anthropological point of view that reality is virtual is this: why do I think that I'm here in this room, in the pupil of my dominant eye? It's because as I move my head around, everything coheres with my mind's eye picture. If I move my hand in front of my face, the motor coordination system -- all the reference frames that the muscles and the brain that control the muscles must know to move my hand smoothly -- is coordinated with the visual reference frame. So I say, "Yes, I'm moving my hand in front of my eyes." But philosophers have known that this is not an ironclad argument. It's a persuasive argument. Persuades me. Persuades everyone else, more or less. Professor Stark designed the optics for the Mars rover Sojourner But I came to the conclusion that virtual reality works because it uses the same mechanisms that make us believe that we're really here in this room and moving around and talking.

You've gone on to apply the knowledge that you've gained from these studies to robots. Replicating in machines what we know through your work about the mind's eye. Tell us about that.

There's a train from a conceptual model to a mathematical model to a mathematical model in a computer that can be made to work and generate solutions, to a physical model, to a robot. In my field I'm interested in movement and vision. A robot is something that moves and has computer vision. And so a little robot can go around the room and pick up Coke cans -- that's a usual task for robots -- or it can go into a nuclear plant and do clean-up. It has these vision and motor capabilities that I've been studying. Some of them are more primitive than human and some of them are more advanced than human capabilities.

I've used models all my life. I think one of my scientific interests has been to develop, to understand, to learn about a new mathematical way of looking at things, and then thinking, "How can I use this mathematical scheme to understand something going on in the brain?" Or in nature, in the evolution of life on the biosphere. So I've been very interested in thinking about robot vision and how I can use robot vision to understand human vision better. I've made a model of robot vision that follows the scan path scheme so I can see it explicitly working out. Also, I can take ideas from human vision to define better robot schemes. We've done both in our laboratory. So the robot vision scheme reinforces our understanding of the human vision. At the same time, what we know about human vision makes us design better robot vision systems.

Next page: The Implications of Science

© Copyright 2000, Regents of the University of California