Clustering-based volume segmentation design

A novel volumetric data clustering work introduced in this paper aim to cluster the volume data and filter out its inherent noise via extracting the data structure and indicating the useless segments. On the basis of classic segmentation algorithms, this research focuses on exploring volume-based segmentation solutions and property-oriented display mechanisms to assist with the decision-making stage involved in associated volume data manipulation works. As the resulting outputs of this design, the occlusion relationships embedded into volumetric space can be precisely oriented in the manner of visualised partition feature(s). This data visualisation process can be accomplished automatically based on the classified information. In addition, a novel manipulation operation can be built via extracting wireframe-based surfaces from the segmentation results. Copyright © 2016 Inderscience Enterprises Ltd.