ISBN: 978-981-11-0008-6 DOI: 10.18178/wcse.2016.06.110
Decomposition of Volumetric Models Based on Distance Transform and K-means Clustering
Abstract— This paper describes a simple decomposition technique for volumetric models. 3D polygonal
models are converted to volumetric models, and these models are decomposed by a standard K-means
clustering technique. Although the standard K-means clustering technique uses randomly generated initial
seeds to determine the starting centroids of the clusters, our technique uses distance maps to determine
unique centroids. Preliminary experiments are conducted on a database of 3D models with various shapes.
Our method shows better decomposition results compared to random decomposition results, since the
technique uses a 3D model’s geometrical features.
Index Terms— K-means, distance transform volumetric data, voxel, decomposition, 3d model.
Motofumi T. Suzuki
The Open University of Japan, JAPAN
Cite: Motofumi T. Suzuki, "Decomposition of Volumetric Models Based on Distance Transform and K-means Clustering," Proceedings of 2016 6th International Workshop on Computer Science and Engineering, pp. 632-635, Tokyo, 17-19 June, 2016.