Computer graphics processing and selective visual display system – Computer graphics processing – Animation
Reexamination Certificate
1998-08-21
2001-12-04
Nguyen, Phu K. (Department: 2772)
Computer graphics processing and selective visual display system
Computer graphics processing
Animation
Reexamination Certificate
active
06326972
ABSTRACT:
BACKGROUND
1. Field of Invention
The present invention relates generally to computer based animation, and more particularly to animating characters in three dimensional (3D) environments. The present invention is also directed to animation techniques for animating scenes with crowds of objects and characters.
2. Background of the Invention
Motion pictures often include scenes of large crowds. For example, battlefields, stadiums, busy urban street scenes, and the like are found in many motion pictures. The most basic conventional film based approach to producing scenes of large crowds was hiring a “cast of thousands” to be the crowd and actually filming the scene. This approach is expensive, time consuming, and often impractical. More sophisticated approaches, for example to fill a stadium, include using a small number of live actors and large numbers of cardboard cutouts of people, or photographing a small number of actors and compositing their images multiple times to create the appearance of a crowd.
Computer-based animation of motion pictures also requires the ability to produce crowd scenes. However, the current technology paradigm of polygon based rendering makes animation of crowds extremely time consuming to produce. Polygon rendering takes a 3D model of a character and covers the surface of the model with a mesh of polygons, each of which has various color, transparency, and/or texture attributes. In conjunction with a lighting model defining sources of lighting and a camera model defining a point of view, the polygons of the model are rendered into an image. Depending on the size and detail of a character, to obtain a realistic, smooth surface of complex character may require about 1,000 polygons are used for each character.
For large crowd scenes, with 1,000 to 10,000 characters, each of a very small size in a final image, it is the silhouettes and shadows of the characters that need to be realistic in order to produce the best animation effects. With a polygon-based animation system, 1,000 characters with 1,000 polygons each would require rendering 1,000,000 polygons for a single frame of animation. At 24 frames per second, a short 1 minute scene would require 1.44 billion polygons rendered. Even at the extremely high speed of 1,000 polygons per second, this 1 minute scene would take about 400 hours to render. Hence, feature length films with many extensive crowd scenes become unworkable with polygon rendering. If the number of polygons per character is reduced in order to speed rendering, then the resulting character silhouettes and shadows become very angular, blocky and unrealistic looking. Thus, polygon rendering, while useful for very small numbers of characters and polygons per frame, becomes unwieldy for crowd scenes. This accounts in part for the lack large crowd scenes of more than a few seconds in current computer animated films.
Accordingly, it is desirable to provide a system and method of rendering 3D character models that can support high speed rendering of large crowds of characters, without sacrificing realistic shadows, silhouettes and movement of the characters.
SUMMARY OF THE INVENTION
The present invention overcomes the limitations of conventional animation systems by providing a fully three-dimensional (3D) stroke-based character animation system and method. The present invention further provides a stroke-based character model, and a user interface particularly adapted for designing the stroke-based character model. Characters may be any object that exhibits movement over time, and which may have portions that move with respect to each other. The present invention may be used to render a single frame of animation, or more preferably scenes of animation containing many frames; in addition, the present invention may be used with non-frame based systems that generate instantaneous images, such as computer generated animation for video games and other real time applications. Accordingly, “frames” and “images” are herein understood to be synonymous.
In one aspect the present invention provides for defining a 3D stroke model for a character using collection of strokes in a 3D environment. The stroke model is directly associated with the 3D geometry of the character so that motion of the character in a 3D environment effects motion of the strokes associated with the character. This allows the positions of strokes of the character in the 3D environment to be computed at any instant in time, and for any specific frame of animation. The stroke position information, in conjunction with a camera and lighting model, is then input into a rendering system which renders an image of the character. In practice a single character can be accurately modeled with a relative few number of strokes, on the order of 15 to 50, in contrast to the approximately 1,000 polygons for a polygon-based character. As a result, a frame of animation including even 10,000 characters needs only about 150,000 to 500,000 strokes rendered, which requires just a few minutes to render. Thus, the present invention makes it practical to use computer animation for scenes of large crowds, even scenes lasting several minutes.
In one embodiment, each stroke of a 3D stroke model is preferably defined by a number of control points which are associated with the character's geometric model. Preferably the control points are associated with joints of the character, so that motion of the joints effects motion of the control points. A stroke passes through 1 or more control points. Each control point has attributes such as size, color, transparency, texture, and the like, that can vary the visual characteristics of the stroke as it passes through the control point. Each control point also has a position in the 3D environment, preferably relative to a local coordinate system of the character. The position of a control point over time depends on a defined relationship to the joint (or other geometry) to which it is assigned, which relation determines how the position is derived over time. Derivation of the control point position may be direct (e.g., where the control point is coincident with a joint) or by interpolation (e.g., where the control point is between two or more joints), or extrapolation. The present invention provides a distinctive user interface that enables the user to view the geometric model of the character, and by direct manipulation establish strokes in association with the geometric model, and view the stroke model in various modes and from a variety of positions. The present invention further enables the creation of stroke variations for each character, so that a given character has different stroke models associated with it, enabling the user to choose one of many different stroke models for rendering the character, as desired. The strokes of a character are preferably stored in a stroke database associated with the character. In addition, the character preferably has a joint database defining its joints and their positions over time for some sequence of animation.
In the method of the present invention, the stroke database of a number of characters is created, so that each character to be included in a frame of animation (or a scene) has a stroke model defined for it. Using a motion system, the positions of the characters in the frame or sequence of frames is determined. This produces information defining the position of each portion (e.g., joint) of each character at each of a number of instances in time corresponding to the frames of the scene. This information is of the joint position within the 3D coordinate systems of each of the characters, and not within a global 3D coordinate system containing all of the characters. The stroke databases of the characters and the character position information is input into a stroke generator, which determines the positions of the strokes of the characters within a global 3D coordinate system, and then translates those positions into a 2D coordinate system for a frame of animation. The result is a collection of 2D stroke data, with colo
Buhler Juan J.
Prasso Luca
Fenwick & West LLP
Nguyen Phu K.
Pacific Data Images, Inc.
LandOfFree
3D stroke-based character modeling suitable for efficiently... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with 3D stroke-based character modeling suitable for efficiently..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and 3D stroke-based character modeling suitable for efficiently... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2600004