基于内存占用的大规模场景的多级ROI分割算法

A Multistage ROI Division Algorithm for Large Scale Scene Based on Memory Cost

随着计算机性能的飞速提升和互联网的蓬勃发展,现有的虚拟现实应用中,所承载的场景越来越大。无论是2D场景或是3D场景,在规模特大后都需要有效组织方能有效率的运行于网络或者个人计算机上。该文提出一种基于内存占用的多级ROI(Region Of Interest,感兴趣区域)分割算法,可以将2D或3D场景有效分割为若干层许多小场景,便于加载和显示。实验证明,该文方法可有效降低内存占用,优化场景显示效率。

With the vigorous development of computer performance and the Internet, scenes in applications of virtual reality existing are getting larger and larger. Both 2D and 3D scenes need management to display effectively. Based on memory cost and ROI theory, we present an algorithm for large scale scene division, which can put a large scale scene into several levels and many small scenes for easy loading and displaying. Experiments show that, this method can effectively reduce the memory cost, and help display efficiency.