![Sample our Computer Science journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5928/TOC-Subject-Sample-2021-Computer-Science.jpg"})
Abstract
An orthogonal BSP (binary space partitioning) tree is a commonly used spatial subdivision data structure for ray-tracing acceleration. While the construction of a BSP tree takes a relatively short time, the efficiency of a traversal algorithm significantly influences the overall rendering time. We propose a new fast traversal algorithm based on statistical evaluation of all possible cases occuring during the traversal of a BSP tree. More frequent cases are handled simply, while less frequent ones are more computationally expensive. The proposed traversal algorithm handles all singularities correctly, and saves between 30% and 50% of traversal time compared with the commonly-known Sung and Arvo algorithms.
