One of the tasks required to display 3D objects is the computation of which surfaces, or parts of surfaces, are visible. Generally this is accomplished by computing the intersections of the various surfaces. Then for each set of intersections, it must be determined which surface is closer to the viewer, and therefore visible.
Two general approaches have been useful in tackling this problem, image space and object space algorithms. These approaches differ in the precision with which they compute the visible surfaces. Image space algorithms determine visible surfaces by examining surfaces at the pixel level whereas object space algorithms directly compare surfaces in the defined space.
This presentation will discuss these two approaches to hidden surface removal and variations to the algorithms. Then, an image space SIMD parallel algorithm will be described in detail.
Colloquia Series page.