Lecture 9: Intro to Ray-Tracing & Accelerating Ray-Scene Intersection (54)
adityasingh7311
For a second I was confused why internal nodes in KD-Trees don't store the min/max bounds of their bounded areas but I think I see why. Since the KD-Tree is a hierarchical structure, the splitting bounds are implicitly passed down to children nodes so there is no need to explicitly consider any value except the current split axis and position.
brywong
What type of data structure do the leaf nodes use to store each object? Is it simply a list of pointers to object references?
LucasArmand
Since we are organizing in 3D space, why do we not use octrees as a data structure for acceleration? My guess would be that we cannot guarantee that a ray stays within an octant of an octree once it enters, so we cannot prune the tree for each ray as well as with a KD-tree
For a second I was confused why internal nodes in KD-Trees don't store the min/max bounds of their bounded areas but I think I see why. Since the KD-Tree is a hierarchical structure, the splitting bounds are implicitly passed down to children nodes so there is no need to explicitly consider any value except the current split axis and position.
What type of data structure do the leaf nodes use to store each object? Is it simply a list of pointers to object references?
Since we are organizing in 3D space, why do we not use octrees as a data structure for acceleration? My guess would be that we cannot guarantee that a ray stays within an octant of an octree once it enters, so we cannot prune the tree for each ray as well as with a KD-tree