Lecture 9: Intro to Ray-Tracing & Accelerating Ray-Scene Intersection (74)
egbenedict
How would we determine what "a few objects" is? Would we just arbitrarily choose some small number, or use some heuristic dependent on other attributes of the scene?
saltyminty
It's usually just an arbitrary number we decide for our stopping condition, and can either be a specific number or dependent on the total number of items we're processing. Not sure if there's any heuristics but the splitting could stop early due to other stopping conditions (ie the slides mentioned max depth).
joeyzhao123
The next slide also provides a bit more information. They listed out 5 as an example for a "few." In addition, when splitting doesn't reduce the expected cost then there is no reason to keep splitting. The "few" probably just has to do with that idea more.
AlexSchedel
I think it is also dependent on the the objects themselves. If you have a lot of objects that are very complex, you might want to limit the bounding boxes, or maybe even split that object itself across multiple bounding boxes, whereas of you have simple objects like in this example, it might be okay to have a bigger bounding box
How would we determine what "a few objects" is? Would we just arbitrarily choose some small number, or use some heuristic dependent on other attributes of the scene?
It's usually just an arbitrary number we decide for our stopping condition, and can either be a specific number or dependent on the total number of items we're processing. Not sure if there's any heuristics but the splitting could stop early due to other stopping conditions (ie the slides mentioned max depth).
The next slide also provides a bit more information. They listed out 5 as an example for a "few." In addition, when splitting doesn't reduce the expected cost then there is no reason to keep splitting. The "few" probably just has to do with that idea more.
I think it is also dependent on the the objects themselves. If you have a lot of objects that are very complex, you might want to limit the bounding boxes, or maybe even split that object itself across multiple bounding boxes, whereas of you have simple objects like in this example, it might be okay to have a bigger bounding box