There are two key elements in defining the problem of visual perceptio
n. The first is that useful information about the world, such as the s
hape, material, illumination, and spatial relationships of objects, is
encrypted in the image. Second, the encryption process, of going from
a description of the world to an image, is not in general reversible.
Any single source of image information is usually ambiguous about its
causes in the scene. Seeing is the process of decoding the image info
rmation. 3-D computer graphics simulates the process of encrypting sce
ne information into the image. By creating images from synthetic scene
s, we can gain insights into the constraints used by the visual system
to decode image information, and we can begin to bridge the gap betwe
en the simple images of the laboratory and complex natural scenes. Com
puter graphics modeling and animation tools provide the means to gener
ate stills and animations that produce strong perceptual interpretatio
ns, yet are theoretically indeterminate. I will describe several illus
ions involving computer renderings and animations that illustrate the
constraints human perception uses to solve ambiguity about material, s
hape, and depth.