Abstract:The kd-tree and Bounding Volume Hierarchy (BVH) are well-known data structures for computing ray-object intersections. Less known is the Constrained Convex Space Partitioning (CCSP), which partitions space and makes the geometric primitives exactly overlap with the boundaries of its cells. Consequently, it is robust against ill-fitting cells that plague methods with axis-aligned cells (kd-tree, BVH) and it permits an efficient, stackless traversal. Within computer graphics, CCSPs have received some attention in both 2D and 3D, but their construction methods were never directly aimed at minimizing their traversal cost -- even having fundamentally opposing goals for Delaunay-type methods. Instead, for an isotropic and translation-invariant ray distribution the traversal cost is minimized by minimizing the weight: the total boundary size of all cells in the structure. We study the 2D case using triangulations as CCSPs and minimize their total edge length using a simulated annealing process that allows for topological changes and varying vertex count. Standard Delaunay-based triangulation techniques show total edge lengths ranging from 10% higher to twice as high as our optimized triangulations for a variety of scenes, with a similar difference in traversal cost when using the triangulations for ray tracing. Compared to a roped kd-tree, our triangulations require less traversal steps for all scenes that we tested and they are robust against the kd-tree's pathological behaviour when geometry becomes misaligned with the world axes. Moreover, the stackless traversal strongly outperforms a BVH, which always requires a top-down descent in the hierarchy. In fact, we show several scenes where the number of traversal operations for our triangulations decreases(!) as the number of geometric primitives $N$ increases, in contrast to the increasing $\log N$ behaviour of a BVH.

Fast ray tracing by ray classification

A Fast Ray-Tracing Algorithm-Based on Adaptive Space Subdivision

A Fast Sah-Based Construction of Octree

Efficient Data Management for Incoherent Ray Tracing.

Incoherent Ray tracing on GPU.

Fast Ray Tracing Procedure Using Space Division with Uniform Rectangular Grid

Multi-level ray tracing algorithm

Fast GPU perspective grid construction and triangle tracing for exhaustive ray tracing of highly coherent rays

Efficient Kd-Tree Construction for Ray Tracing Using Ray Distribution Sampling

Realtime Ray Tracing on a Hibrid Parallel Architecture

KD-TREE BASED FAST RAY TRACING FOR RCS PREDICTION

Fast Physical Optics Integrated Based on Vulkan Ray Tracing

Using Hardware Ray Transforms to Accelerate Ray/Primitive Intersections for Long, Thin Primitive Types.

Stackless Ray-Object Intersections Using Approximate Minimum Weight Triangulations: Results in 2D That Outperform Roped KD-Trees (And Massively Outperform BVHs)

Faster Ray Tracing through Hierarchy Cut Code

3D Tetrahedron Ray Tracing Algorithm

A High-Resolution Compression Scheme for Ray Tracing Subdivision Surfaces with Displacement

A Framework for Quantum Ray Tracing

An Efficient Rejection Test for Ray/Triangle Mesh Intersection

3D Gaussian Ray Tracing: Fast Tracing of Particle Scenes