Visualizing Query Traversals Over Bounding Volume Hierarchies Using Treemaps

Abhishek Madan, Carolina Nobre

Room: 104

2023-10-25T03:00:00ZGMT-0600Change your timezone on the schedule page
2023-10-25T03:00:00Z
Exemplar figure, described by caption below
An annotated screenshot of our visualization, containing a 3D viewport, a zoomable treemap, and a pixel grid. All three views work together to show the structure of a bounding volume hierarchy (BVH) with coordination between the viewport and treemap, view statistics of ray intersection queries with the BVH, and view individual query traces and results.
Fast forward
Full Video
Keywords

bounding volume hierarchies, treemaps

Abstract

Bounding volume hierarchies (BVHs) are one of the most common spatial data structures in computer graphics. Visualizing ray intersections in these data structures is challenging due to the large number of queries in typical image rendering workloads, the spatial clutter induced by superimposing the tree in a 3D viewport, and the strong tendency of these queries to visit several tree leaves, all of which add a very high dimensionality to the data being visualized. We present a new technique for visualizing ray intersection traversals on BVHs over triangle meshes. Unlike previous approaches which display aggregate traversal costs using a heatmap over the rendered image, we display detailed traversal information about individual queries, using a 3D view of the mesh, a treemap of the BVH, and synchronized highlighting between the two views, along with a pixel grid to select a ray intersection query to view. We demonstrate how this technique elucidates traversal dynamics and tree construction properties, which makes it possible to easily spot algorithmic improvements in these two categories.