Lecture 9: Intro to Ray-Tracing & Accelerating Ray-Scene Intersection (57)
mylinhvu11
I understand the pros of using ideal KD-Tree Pre-Processing, but how can we code an algorithm that can analyze what can reduce the expected cost of ray intersection? What variables and factors do we need to be able to determine this? Or are these external factors we must decide while we are coding for the parameters?
yzliu567
The accurate expected cost of ray intersection is hard to model and compute, but we have some approximate methods. For example, we can assume that the intersection probability is proportional to the surface area. By also estimating the cost of different basic operations during the KD-tree search, we can apporximately compute the expected cost. See later slides for details.
I understand the pros of using ideal KD-Tree Pre-Processing, but how can we code an algorithm that can analyze what can reduce the expected cost of ray intersection? What variables and factors do we need to be able to determine this? Or are these external factors we must decide while we are coding for the parameters?
The accurate expected cost of ray intersection is hard to model and compute, but we have some approximate methods. For example, we can assume that the intersection probability is proportional to the surface area. By also estimating the cost of different basic operations during the KD-tree search, we can apporximately compute the expected cost. See later slides for details.