Abstract

Rendering of 3D scenes with a large number of objects is computationally intensive. Occlusion culling methods are used to improve rendering performance by reducing the number of objects to be processed on the GPU. In this paper, we consider occlusion culling methods that employ spatial and temporal coherence of visibility. Some of the most efficient methods are based on hardware occlusion queries. However, recording of occlusion queries and receiving their results for a large number of objects can take considerable time. We propose an algorithm that removes occlusion query transmission overhead by performing occlusion checks on the CPU against a depth buffer downloaded from the GPU. The proposed algorithm is compared with widespread algorithm based on hardware occlusion queries. It improves rendering performance of 3D scenes with a large number of objects by reducing the number of transmitted rendering commands.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call