节点文献
基于插值矢量量化的地形金字塔纹理压缩
Terrain Pyramid Texture Compression using Interpolative Vector Quantization
【摘要】 矢量量化因其压缩比大,解码简单效率高成为实时绘制系统中大规模纹理压缩主要使用的方法。文献[1]介绍的基于矢量量化方法针对纹理金字塔压缩取得了很大的成功。在此基础上使用一个特征值数据本,一个残差矢量码本和索引表对金字塔纹理的各层进行压缩,在保持实时性的同时提高了压缩比。使用了改进的SOFM码本生成算法,实验表明其码本质量优于传统的码本生成算法。
【Abstract】 Vector quantization(VQ) method becomes the main method used in real time visualization system because of its high compression ratio and fast decoding speed.Paper [1] introducs an extension to the basic VQ technique for com-pressing mipmaps.This paper describes the method of using one representative value databook,one residual codebook and one index map to compress all levels of the pyramid texture.It increases the compression radio.It presents the SOFM algorithm.The result shows that the quality is higher than the traditional codebook-generated algorithm.
【Key words】 texture compression; pyramid texture; vector quantization; SOFM;
- 【文献出处】 计算机工程与应用 ,Computer Engineering and Applications , 编辑部邮箱 ,2006年22期
- 【分类号】TP391.41
- 【被引频次】1
- 【下载频次】100