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.

Abstract: A method, apparatus, and software product interpolates inbetween shapes of a two- or three-dimensional geometric model between a sequence of key poses of the model using a new type of parametric curve with shape parameters for providing independent control over the properties of the curve. The shape parameters provide the animator with full control over the behavior of the interpolated shapes between one key pose to the next. More specifically, these shape parameters independently control the direction, absolute magnitudes, and relative magnitudes of the incoming and outgoing tangent vectors at each key pose.

Abstract: A method, apparatus, and software product generates deformation weights for application of a deformation function on a geometric model. The user, working directly on the points of the geometric model, including its surface, defines a deformation area having a focal region fully affected by the deformation and a falloff region only partially affected by the deformation. Deformation weights are then automatically calculated for all points in the regions affected by the deformation, using a technique that creates a smooth and natural decrease in weights from the focal region to the falloff region. Several options are available to the user to control and monitor deformation weight calculations and weighted blending, and to visualize the deformation weights themselves.

Abstract: A method apparatus, and software product generates raster-based complete outlines of raster-based rendered images based on 3D models. Preliminary outlines of a rendered image are created from the subregion identifiers, position data, and orientation data from the rendered image database. These preliminary outlines are then combined and refined to generate the final image outline. Subregion identifiers identify various subregions or materials in the image. Position data describes a distance between a pixel and a reference point. The orientation data describes the orientation of the underlying models at each pixel in the image. In one embodiment, orientation data describes variations in orientation at a given pixel with respect to some number of nearby pixels.

Abstract: Method and apparatus are disclosed for generating mattes of film, video, or other visual image elements, in which the operator may specify soft-edge regions for the extended visible fall-off of certain difficult areas such as hair, clouds, shadows, and blurred objects. The elements may be either still or in motion, and automatic interpolation of keyframe regions may be used to simplify the generation of in-between frames to create traveling soft-edge mattes.