Painting surface characteristics

Abstract

Surface illumination proceeds according to the strict rules of the reeection model, the light characteristics and the geometry of the scene. To get a given illumination eeect on a surface in a axed geometry, a user must determine the surface characteristics that will produce this eeect. This process is part of what is called inverse illumination. In most current modeling systems, the user must rely only on intuition to perform this inverse illumination, which can lead to many modeling/rendering cycles to achieve a satisfactory result. In this paper, we are concerned with the case where lighting eeects are not merely a consequence of the geometry, but rather part of the design. We therefore concentrate our eeorts on controlling the surface character-istics and present a tool which can reduce considerably the problem of inverse illumination by using a painting paradigm. An interactive system is provided where the user simply applies color points on a surface. The system then attempts in near real time to ond the best values for the sur-face characteristics such that the points will retain their assigned color in the nal rendering. Depending on the number of constraints (color points) given by the user, the solution is presented as a non-linear con-strained optimization and a constrained weighted least-square etting. We apply our solution to a simple illumination model using ambient, diiuse and specular components to illustrate our approach.