AbstractA computer's rendering capacity cannot handle the increased amount of data when rendering high‐quality, three‐dimensional computer graphic images. This results in problems such as rendering not being possible depending on the subject image, or the rendering speed dropping too far. To resolve these problems, we propose a method for generating a new hierarchical three‐dimensional data structure based on recursive face cluster partitioning, and a high‐speed rendering data extraction algorithm using that data structure. In the proposed algorithm, local resolution control using the approximating polygon of a face cluster, and precise invisible surface culling using the bounding volume of the face cluster can be executed in parallel. High‐speed rendering can be performed with little deterioration in quality, even for a large amount of data. The effectiveness of this algorithm was confirmed with a local resolution control experiment and a large‐scale three‐dimensional model rendering experiment. © 2007 Wiley Periodicals, Inc. Syst Comp Jpn, 38(8): 32– 43, 2007; Published online in Wiley InterScience (www.interscience. wiley.com). DOI 10.1002/scj.20581
Read full abstract