Transparent visualization of large-scale and complex polygon meshes using a stochastic point-based rendering method

Efficient and reliable transparent visualization of large and complex surface data is important in many applications, including the visualization of medical data. Among the many methods, depth-peeling is an efficient solution for rendering polygon meshes transparently; however, the computational cost is related to the number of passes through the peeled geometry. Therefore, the depth-peeling method cannot efficiently render large and complex polygon meshes that require many passes through the geometry. In this paper, we solve this problem using a stochastic point-based rendering (SPBR) method. We use two types of large-scale and complex polygon meshes that have 5,744,376 and 4,800,644 triangles in the experiments. The rendering speed of our method is 3-5 times faster than that of the depth-peeling method for an image resolution of 512 × 512. Additionally, the SPBR method has an inherent visual effect: the outlines of target polygon meshes can be automatically enhanced. This makes the results of the SPBR method more vivid and comprehensible than those of the depth-peeling method.