基于GPU加速的轮廓毛发绘制技术

GPU-based silhouette fur rendering

在采用层状纹理切片来表达真实感毛发的方法中,模型轮廓处毛发切片的计算与生成需要耗费很多时间,影响绘制效率。针对此问题,提出一种利用图形处理器(GPU)可编程能力的轮廓毛发快速生成技术。该技术通过为顶点设计一种"边信息"的数据结构,将轮廓边的检测及轮廓毛发切片的生成完全转移到了GPU中进行,从而大大加速了轮廓毛发的处理,提高了真实感毛发的绘制效率。

In those methods that represent furry surfaces with multi-layer textured slices,the computation and generation of silhouette fur always spend much time,affecting rendering efficiency greatly.To solve the problem,this paper presents an approach that can generate silhouette fur very fast based on the programmable ability of Graphic Process Units （GPU）.In the approach,by using an ＂edge information＂ data structure,the silhouette-edge detection and silhouette fur slice generation are wholly transformed into GPU ,hence greatly shortening the time spending on silhouette fur and improving the rendering efficiency.