In our previous Three Dimensional (3D) anthropometric shape clustering study, block-division technique is adopted. The objective of this study was to examine the sensitivity of clustering results on block-division. Such a block-division technique means to divide each 3D surface into a predefined number of blocks. Then by using a block-distance measure, each surface is converted into a block-distance based vector. Finally, k-means clustering is performed on the vectors to segment a population into several groups. Totally 447 3D head samples have been analyzed in the case study. The influence of block division number on clustering was evaluated by using One-way ANOVA. No significant difference was found between the three block division alternatives. This means the adopted method is robust to block division. © 2009 Springer Berlin Heidelberg.
CITATION STYLE
Niu, J., Li, Z., & Xu, S. (2009). Block Division for 3D Head Shape Clustering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5620 LNCS, pp. 64–71). https://doi.org/10.1007/978-3-642-02809-0_8
Mendeley helps you to discover research relevant for your work.