PROJECT TITLE:

Voronoi Cell-Based Clustering Using a Kernel Support

ABSTRACT:

Support-primarily based clustering using kernels suffers from serious computational limitations inherent in many kernel ways when applied to terribly giant-scale problems despite its ability to identify clusters with advanced shapes. In this paper, we tend to propose a novel clustering algorithm known as Voronoi cell-based mostly clustering to expedite support-primarily based clustering using kernels. In contrast to previous studies, including the basin cell-based mostly technique, the proposed method achieves computational efficiency in each the training part to construct a support estimate using sampled knowledge to reduce the analysis of kernels and therefore the labeling part to assign a cluster label on every knowledge point nearest its representative purpose. The performance superiority of the proposed method over the other basin cell-primarily based methods in terms of computational time and storage potency is verified by various experiments using benchmark sets and in real applications to image segmentation.