Image occlusion processing using both object bounding and object bounded volumes

When rendering a scene 1 including a complex object 3 made up of many individual primitives, rather than processing each primitive in turn, a bounding volume 4 which surrounds the complex object 3 is generated and the scene 1 is then processed using the bounding volume 4 in place of the actual object primitives. The claimed method identifies a scene object 3, generates a first bounding (i.e. object covering, surrounding or enclosing) volume 4 representing the object and assesses whether the bounding volume is at least partially occluded (overlapped) by other scene objects, e.g. feature 2; the method further comprises generating an object bounded (i.e. object covered, surrounded or enclosed) volume (70, Figure 11) and assessing whether this bounded volume, also representing the object, will at least partially occlude (cover, obscure) other scene objects; an object processing determination is made accordingly. Further (subdividing) bounding volumes may be used (Figures 8-10). Depth (z) only rendering may be performed, based on bounding or bounded volumes, or corresponding primitives, being tagged with an identifier.