We propose a new method for efficient out-of-core rendering of complex 3D scenes. The method consists of two main phases: a new out-of-core preprocessing and a run-time, view-dependent rendering of the scene. In the preprocessing, wo partition the scene with a spatial octree and build an on-disk multi-resolution hierarchical representation for the scene. Then at run-time, the multi-threaded technology is used. The prefetching thread predicts the motion of the viewer and prefetches the scene that would be seen next. The rendering thread traverses the octree hierarchy and renders the scene by view-dependent selective refinement. Experimental results show that our system can render large complex 3D scenes of tens of millions of triangles at interactive rates on commodity PC with fine image quality and minimal popping artifacts. © Springer-Verlag Berlin Heidelberg 2006.
CITATION STYLE
Gao, Y., Deng, B., Wei, Y., & Wu, L. (2006). Efficient out-of-core rendering of complex 3D scenes. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3942 LNCS, pp. 883–892). https://doi.org/10.1007/11736639_108
Mendeley helps you to discover research relevant for your work.