Semi-automatic segmentation of 3d point clouds skeleton without explicit computation for critical points

Kok, Why Ng and Abdullah Junaidi, and Sew, Lai Ng (2012) Semi-automatic segmentation of 3d point clouds skeleton without explicit computation for critical points. In: PRICAI'12 Proceedings of the 12th Pacific Rim international conference on Trends in Artificial Intelligence. Springer-Verlag Berlin, pp. 783-788. ISBN 978-3-642-32694-3

[img] Text
chp%3A10.1007%2F978-3-642-32695-0_73(1).pdf - Published Version
Restricted to Repository staff only

Download (485kB)
Official URL: http://dl.acm.org/citation.cfm?id=2402351

Abstract

Segmentation of 3D point clouds is vigorously discussed in recent years. Many existing techniques pre-process the data to identify critical points for meaningful features. Very often, the critical points are predicted at curvature objects and this does not always generate promising outcome. This paper proposes to segment the point clouds based on its skeleton as this robustly reflects the global shape of an object. The skeleton is constructed via Laplacian-based contraction method. Spherical approach is applied along the skeleton to segment the point clouds. The entire process is automatic. Only moderate user-input is applied to align the skeleton to the object. The output of the proposed method is to be compared with another popular segmentation method with critical points input. The result shows that the proposed method generates more accurate segmented features.

Item Type: Book Section
Subjects: Q Science > QA Mathematics > QA75.5-76.95 Electronic computers. Computer science
Divisions: Faculty of Computing and Informatics (FCI)
Depositing User: Ms Suzilawati Abu Samah
Date Deposited: 01 Nov 2013 03:21
Last Modified: 01 Nov 2013 05:07
URI: http://shdl.mmu.edu.my/id/eprint/4339

Actions (login required)

View Item View Item