An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation
Three-dimensional (3D) model segmentation has received tremendous attention in recent years, to partition model mesh into meaningful sub-meshes. Hitherto, there is no robust and consistent segmentation method to overcome the problems of under- and over-segmentation for the meaningful components. Man...
Saved in:
Main Author: | |
---|---|
Format: | Thesis |
Published: |
2016
|
Subjects: | |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
my-mmu-ep.7147 |
---|---|
record_format |
uketd_dc |
spelling |
my-mmu-ep.71472018-03-16T16:47:13Z An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation 2016-10 Ng, Kok Why TA1501-1820 Applied optics. Photonics Three-dimensional (3D) model segmentation has received tremendous attention in recent years, to partition model mesh into meaningful sub-meshes. Hitherto, there is no robust and consistent segmentation method to overcome the problems of under- and over-segmentation for the meaningful components. Many existing methods require either user-input seeding for the number of segments or to apply minima rules to approximate the meaningful components. Some methods excel only in a narrow range of models. Their methods are vague, sensitive to model shape (unstable) and tedious (duplicated processes). Slinky-based segmentation (SBS) with improved skeleton method is proposed in this thesis to automatically and consistently identify meaningful features of a model. The method is robust on any input model shape. The algorithm begins with voxelization and surface-reconstruction on the input model to get rid of the irregular meshes. Laplacian-based Contraction method is adapted to shrink the model into triangular skeleton. 2016-10 Thesis http://shdl.mmu.edu.my/7147/ http://library.mmu.edu.my/diglib/onlinedb/dig_lib.php phd doctoral Multimedia University Faculty of Computing and Informatics |
institution |
Multimedia University |
collection |
MMU Institutional Repository |
topic |
TA1501-1820 Applied optics Photonics |
spellingShingle |
TA1501-1820 Applied optics Photonics Ng, Kok Why An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation |
description |
Three-dimensional (3D) model segmentation has received tremendous attention in recent years, to partition model mesh into meaningful sub-meshes. Hitherto, there is no robust and consistent segmentation method to overcome the problems of under- and over-segmentation for the meaningful components. Many existing methods require either user-input seeding for the number of segments or to apply minima rules to approximate the meaningful components. Some methods excel only in a narrow range of models. Their methods are vague, sensitive to model shape (unstable) and tedious (duplicated processes). Slinky-based segmentation (SBS) with improved skeleton method is proposed in this thesis to automatically and consistently identify meaningful features of a model. The method is robust on any input model shape. The algorithm begins with voxelization and surface-reconstruction on the input model to get rid of the irregular meshes. Laplacian-based Contraction method is adapted to shrink the model into triangular skeleton. |
format |
Thesis |
qualification_name |
Doctor of Philosophy (PhD.) |
qualification_level |
Doctorate |
author |
Ng, Kok Why |
author_facet |
Ng, Kok Why |
author_sort |
Ng, Kok Why |
title |
An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation |
title_short |
An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation |
title_full |
An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation |
title_fullStr |
An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation |
title_full_unstemmed |
An Improved Skeletion-Based Technique For Three-Dimensional Model Segmentation |
title_sort |
improved skeletion-based technique for three-dimensional model segmentation |
granting_institution |
Multimedia University |
granting_department |
Faculty of Computing and Informatics |
publishDate |
2016 |
_version_ |
1747829653344616448 |