CAREER: Light: A Fundamental Primitive for Computer Graphics

点击翻译按钮获取中文摘要

The process of creating a virtual environment or object in computer graphics typically begins with modeling the geometric and surface attributes of the objects in the environment along with any lights. Great effort has been expended to develop computer aided design systems that allow the specification of complex geometry and material attributes. Similarly, a great deal of work has been undertaken to produce systems that simulate the propagation of light through virtual environments to create realistic images. Despite these efforts, it has remained difficult or impossible to recreate much of the complex geometry and subtle lighting effects found in the real world. The modeling problem can potentially be bypassed by capturing the geometry and material properties of objects directly from the real world using some combination of cameras, structured light, range finders, and mechanical sensing devices. When successful, the results can be fed into a rendering program to create images of real objects and scenes. Unfortunately, these systems are still unable to completely capture small details in geometry and material properties. Existing rendering methods also continue to be limited in their capability to faithfully reproduce real world illumination, even if given accurate geometric models. This project will develop techniques for using a representation of light (radiance) as a basic primitive for computer graphics. This primitive will augment the traditional geometry-based primitives used in computer graphics. This framework will provide a unified method for encoding the complete appearance of a complex model, and allow the rerendering of an object at speeds independent of the its geometric or lighting complexity. Moreover, this framework will allow one to use a video camera to sample the radiance information of a real world scene or environment. This sampled information can then be organized properly to allow for redisplay of the scene interactively from any virtual camera at any viewpoint. The research is fundamental in that it explores an entirely novel primitive for computer graphics. This primitive is one of the most natural ways to represent the complicated visual information present in a scene. The research will facilitate the interactive viewing of captured real world scenes. No other known method has the ability to directly represent the information captured by a video camera, and use this information for interactive generation of new realistic images. The success of this project will have impact in such diverse fields as simulation, interactive virtual environments computer education, the visual archival of objects, and the transmittance of the appearance of objects in 3D catalogs.