In many applications, iso-surface is the primary method for visualizing the structure of 3D density maps. We consider a common scenario where the user views the iso-surfaces from a distance and varies the level associated with the iso-surface as well as the view direction to gain a sense of the general 3D structure of the density map. For many types of density data, the iso-surfaces associated with a particular threshold may be nested and never visible during this type of viewing. In this paper, we discuss a simple, conservative culling method that avoids the generation of interior portions of iso-surfaces at the contouring stage. Unlike existing methods that perform culling based on the current view direction, our culling is performed once for all views and requires no additional computation as the view changes. By pre-computing a single visibility map, culling is done at any iso-value with little overhead in contouring. We demonstrate the effectiveness of the algorithm on a range of bio-medical data and discuss a practical application in online visualization.

中文翻译：

---

用于远场查看等值面的 3D 密度图的与视图无关的轮廓剔除

在许多应用中，等值面是可视化 3D 密度图结构的主要方法。我们考虑一个常见的场景，即用户从远处查看等值面并改变与等值面相关的级别以及视图方向，以了解密度图的一般 3D 结构。对于许多类型的密度数据，与特定阈值相关联的等值面可能是嵌套的，并且在此类查看期间从不可见。在本文中，我们讨论了一种简单、保守的剔除方法，该方法可避免在轮廓绘制阶段生成等值面的内部部分。与根据当前视图方向执行剔除的现有方法不同，我们的剔除对所有视图执行一次，并且随着视图的变化不需要额外的计算。通过预先计算单个可见性图，可以在任何等值处进行剔除，而轮廓绘制的开销很小。我们证明了该算法对一系列生物医学数据的有效性，并讨论了在线可视化中的实际应用。