Abstract | ||
---|---|---|
In this paper a novel clustering algorithm is proposed, namely Vari- ational Multilevel Mesh Clustering (VMLC). The algorithm incor- porates the advantages of both hierarchical and variational (Lloyd) algorithms, i.e. the initial number of seeds is not predefine d and on each level the obtained clustering configuration is quasi -optimal. The algorithm performs a complete mesh analysis regarding the un- derlying energy functional. Thus, an optimized multilevel cluster- ing is built. The first benefit of this approach is that it resolves the inher - ent problems of variational algorithms, for which the result and the convergence is strictly related to the initial number and se lection of seeds. On the other hand, the greedy nature of hierarchical ap- proaches, i.e. the non-optimal shape of the clusters in the h ierarchy, is solved. We present an optimized implementation based on an in- cremental data structure. We demonstrate the generic nature of our approach by applying it for the generation of optimized multilevel Centroidal Voronoi Diagrams and planar mesh approximation. Index Terms: I.3.5 ( Computer Graphics): Computational Geom- etry and Object Modeling —Geometric algorithms; I.5.3 ( Pattern Recognition): Clustering—Algorithms |
Year | DOI | Venue |
---|---|---|
2008 | 10.1109/SMI.2008.4547971 | Shape Modeling International |
Keywords | Field | DocType |
approximation theory,computational geometry,mesh generation,VMLC,centroidal Voronoi diagram,planar mesh approximation,quasioptimal clustering configuration,variational multilevel mesh clustering,I.3.5 [ Computer Graphics]: Computational Geometry and Object Modeling—Geometric algorithms,I.5.3 [ Pattern Recognition]: Clustering—Algorithms | Canopy clustering algorithm,Mathematical optimization,CURE data clustering algorithm,Correlation clustering,Mesh analysis,Constrained clustering,Voronoi diagram,Cluster analysis,Mesh generation,Mathematics | Conference |
ISBN | Citations | PageRank |
978-1-4244-2261-6 | 3 | 0.41 |
References | Authors | |
8 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Iurie Chiosa | 1 | 9 | 0.87 |
Andreas Kolb | 2 | 783 | 71.76 |