Output Sensitive Extraction of Silhouettes from Polygonal Geometry

clearly be used in nding the new set of cells and/or edges that needs to be visited in the grid, building the solution of the next silhouette query based on the solution of the previous query. (a) (b) Figure 9: A dinner scene. In (a), the polygonal model and in (b), the silhouette and boundary. bounded by O(n) and hence is in the same order of the magnitude of the input, large models can take full advantage of the real memory of contemporary workstations. Because large models is where the presented algorithm shines, even a decrease by a constant factor of the memory use is signiicant, and eeort should be invested to minimize this memory overhead. The use of this eecient silhouette extraction algorithm should be examined in various other applications from hidden line removal, through general swept volume of object O, to collision detection 7]. While in this work, we have selected a circumscribing cube for simplicity, one should expect somewhat better behavior if a tetrahedra will be used, minimizing the number of faces that need to be processed. While each face is now somewhat more complex, the over whole cost should be evaluated, seeking the best convex circumscribing polyhedra. Furthermore, because objects have diierent silhouette complexity from diierent views, it is also worth while to ask the question of the best circumscribing convex polyhedra, for some special objects. Because of the way the arcs are centrally projected onto the plane of the circumscribing polyhedra, one could also have chosen to subdivide the planes in a radial way. This subdivision scheme should improve our results since the associated segments are expected to be more densely populated in the centers of the six planes, due to the behaviour of the central projection, while making the traversal of cells for a given L i (~ V) more complex. This density should be computed to follow the Riemannian metric of the central projection. In addition, the innuence of the resolution of the grid on the query time, considering both the shape of the model and the number of silhouette edges, is yet to be explored as well, for optimality. Finally, we have not exploited the fact that the stabbing line is continuously moving. This coherency (b) Figure 7: The Utah Teapot. In (a), the polygonal model and in (b), the silhouette and the boundary are shown. (a) (b) Figure 8: A B58 …