A Practical Analysis of Clustering Strategies for Hierarchical Radiosity

Abstract

The calculation of radiant energy balance in complex scenes has been made possible by hierarchical radiosity methods based on clustering mechanisms. Although clustering offers an elegant theoretical solution by reducing the asymptotic complexity of the algorithm, its practical use raises many difficulties, and may result in image artifacts or unexpected behavior. This paper proposes a detailed analysis of the expectations placed on clustering and compares the relative merits of existing, as well as newly introduced, clustering algorithms. This comparison starts from the precise definition of various clustering strategies based on a taxonomy of data structures and construction algorithms, and proceeds to an experimental study of the clustering behavior for real‐world scenes. Interestingly, we observe that for some scenes light is difficult to simulate even with clustering. Our results lead to a series of observations characterizing the adequacy of clustering methods for meeting such diverse goals as progressive solution improvement, efficient ray casting acceleration, and faithful representation of object density for approximate visibility calculations.

Wavelet Radiosity on Arbitrary Planar Surfaces

“…First, our algorithm could be used to group together neighbouring coplanar patches in a natural way. This would help the clustering strategy [14] and give a more accurate result. Second, we would like to integrate our algorithm with face-clustering, bringing multi-wavelets into face-clusters.…”

“…The cluster receives radiosity and distributes it to the patches that it contains. On the other hand, current clustering strategies are behaving poorly in scenes with many small patches located close to each other [14]. It would probably be more efficient to apply clustering to the original planar surfaces instead of applying it to the result of the tessellation.…”

An efficient instantiation algorithm for simulating radiant energy transfer in plant models

“…Its efficiency toward hierarchical radiosity is then improved by inserting new levels of (non instantiable) clusters, using a constrained clusterizer [Hasenfratz et al 1999]. Besides, self-similarity occurs in vegetation scenes at multiple scales including groups of plants of various sizes, and there is no reason to limit the instantiable hierarchy to the level of the plants themselves, as soon as we manage to compute (or predict) their phase functions.…”

Hierarchical Higher Order Face Cluster Radiosity for Global Illumination Walkthroughs of Complex Non‐Diffuse Environments

“…Volume clusteringmethods 3 , 2 , 4 combat this problem by grouping inputpatches into volume clusters. Handling the light incident ona cluster is, however, a difficult problem and all presentedsolutions are more suitable to handling unorganized sets ofpolygons than highly tessellated models 5 , 6 , 7 . It is difficultto obtain continuously shaded surfaces, since interpolatingscalar irradiances across volumes does not lead to good resultsbecause of the varying orientations of surfaces withinthe cluster 6 .…”

“…Handling the light incident ona cluster is, however, a difficult problem and all presentedsolutions are more suitable to handling unorganized sets ofpolygons than highly tessellated models 5 , 6 , 7 . It is difficultto obtain continuously shaded surfaces, since interpolatingscalar irradiances across volumes does not lead to good resultsbecause of the varying orientations of surfaces withinthe cluster 6 . At the same time, pushing irradiance to leaveson‐the‐fly 8 , 3 , 9 makes it difficult to construct higher orderrepresentations of polygon irradiances, makes the methodcomplexity dependent on input model size, and drasticallyreduces the memory locality of the solution phase.…”