3d stroke-based character modeling suitable for efficiently rendering large crowds

Abstract

A system, method, and software product provide an efficient mechanism for rendering characters for computer based animation. The operation of the system is particularly well suited to the rendering of scenes of large numbers of characters, such as scenes of large crowds. The system uses a stroke based character model for a character, which is a collection of strokes, defined in a three dimensional coordinate system of a character, and associated with a geometric model of the character. This stroke model of the character and its association to the geometric model enables the strokes to move or change position over time as the character moves. The stroke model reduces the amount of data needed to accurately represent the character, particularly its shadows and silhouettes. Using the stroke models for many instances of characters within a frame or sequence of frames of animation, the system determines the positions of the characters in a global coordinate system in the frame, and then determines the respective positions of the stroke models of the characters, also in the global coordinate system. This collection of strokes is then input into a renderer, along with a camera model and a lighting model in order to render a final image of a frame or frame sequence. A preferred renderer is a particle renderer which draws each stroke through a series of control points which define the stroke.