This paper presents a new approach for efficient 3Dudrendering on mobile devices, where selective renderingudcan be achieved with the help of 3D scene annotation.udBy taking advantage of first person environments inudmost 3D applications, we are able to annotate theudflooring details of the 3D space. This allows 3Dudenvironments to be interfaced using a higher level viewudof objects. With the higher level of sceneudunderstanding, it is possible to determine which 3Dudobjects are not required for loading or renderingudbased on the viewer’s location and its surroundingudconstraints.
展开▼