双向纹理函数的有效压缩与绘制  

An Algorithm for Efficient Compression and Rendering of BTF

作　　者：吴向阳[1] 龚怿[2] 彭群生[2] 王毅刚[1] 

机构地区：[1]杭州电子科技大学图形图像研究所,杭州310018 [2]浙江大学CAD&CG国家重点实验室,杭州310027

出　　处：《计算机学报》2006年第12期2201-2207,共7页Chinese Journal of Computers

基　　金：国家"九七三"重点基础研究发展规划项目基金(2002CB312101);国家自然科学基金(60303028);浙江省自然科学基金(R603046)资助~~

摘　　要：提出了一种有效的BTF压缩方法:对简单BTF样本区域用一个统一的低频函数模拟整个区域的整体光照效果,再对每个像素生成一个高频函数来表现细节;对复杂样本区域则首先进行像素聚类,再在各个像素类内对视线采样方向自适应地聚类,在各个视线类内分别拟合一个低频函数,并求像素在各个视线内的高频函数,利用低频函数与高频函数的和重建了像素在相应视线类内的视觉效果.与局部PCA方法的对比表明,该文的算法取得了较高的压缩效率和更快的绘制速度,能实现交互的软件绘制和实时硬件绘制.This paper presents an efficient algorithm for compression and rendering of BTFs. For simple BTF sample sets, a single low frequency illumination function defined on the whole BTF and a high frequency function for each pixel are proposed to recover the appearance of the BTF. For complicated texture surfaces, A clustering process is first applied to all pixels, then in each pixel cluster, all sampled views are clustered too. For every view cluster, a low frequency illumination model is fitted, and the diverse high frequency illumination details at each pixel regarding different view directions is calculated. The experiments on a simple BTF set and three complicated sets demonstrate high compression rates and interactive rendering speed by software and realtime performance by hardware.

关 键 词：双向纹理函数(BTF) 双向反射率分布函数(BRDF) GPU 压缩 绘制 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]