Published in SeDICI: Techniques for an Image Based Occlusion Culling Engine


In this work we present several techniques applied to implement an Image Based Software Occlusion Culling Engine to increase the speed of rendering general scenes with high depth complexity. This conservative culling method is based on a tiled Occlusion Map that is updated only when needed, deferring and even avoiding the expensive per pixel rasterization process. We show how the tiles become a useful way to increase the speed of visibility tests. Finally we describe how different parts of the engine were parallelized using OpenMP directives and SIMD instructions.


Leandro R. Barbagallo, Matias N. Leone, Mariano M. Banquiero, Diego Agromayor, Andres Bursztyn

Presented at

X Workshop on Computer Graphics, Images and Visualization (WCGIV) Argentine Congress on Computer Sciences (CACIC) CACIC 2012 WCGIV

Source Code

Source code ISOCE (Image Space Occlusion Culling Engine) Source Code Here