The goal of 3D photography is the inverse of of computer aided
manufacturing: given a physical object, such as a clay model of a car or a human figure, create a computer model capturing its shape and appearance. In mathematical / statistical terms modeling shape means finding a 2-dimensional manifold approximating a collection of points in 3-space. Modeling appearance means estimating a function on a 2-dimensional manifold from sampled data. The range of the target function can be a subset of euclidean space (the RGB coordinates of
"color") of it can itself be a function space (such as the space of "lumispheres"). We will describe some of the work of the 3D photography group at the University of Washington, with special emphasis on the use of multi-resolution methods.
Copyright. All Rights Reserved.