histogram of the texture look like in an image taken from a given viewpoint?
This paper considers three-dimensional textures composed of opaque planar texels uniformly distributed over a volume 3. (Recovery) What geometric properties of the texels of space. When such a texture is viewed from a distance (orthocan be recovered from single images? graphic projection), the probability of seeing through a giventhickness volume of the texture depends only on the spatial In these papers we regard textures such as snow and tree density of the texels and on their distribution of areas and crowns as composed of planar texels (flakes or leaves) slants (of their normals away from the viewing direction); it randomly distributed in space. In the tree case we ignore does not depend on texel shape or on the distribution of tilt the branches, except for their effect on the leaf orientadirections. We evaluate the probability for various texel orientations.
tion models including uniformly distributed orientations, paral-
How ''transparent'' is a 3D texture that occupies a vollel texels (e.g., phototropic leaves), ''drooping'' texels, and texels ume of given thickness t and is composed of opaque planar composed of ''leaves'' whose stems conform to a standard tree texels? In other words, what is the probability that a viewbranching model. In real scenes containing falling disks (''snowing ray will pass completely through the volume without flakes'') we find that the disks have about the same average being blocked by a texel? (More generally: if an object is slant as texels whose distribution of orientations is uniform; located at depth t within the texture, what is the probability thus the uniform orientation model should be the appropriate one for predicting visibility through a snowstorm.