Evaluation of Coherent Hierarchical Culling Revisited with Varied Parameters

Information

• Publication Type: Bachelor Thesis
• Workgroup(s)/Project(s):
  • Rendering and Modeling
• Date: June 2019
• Date (Start): 12. October 2015
• Date (End): 30. June 2019
• Matrikelnummer: 01025543
• First Supervisor: Michael Wimmer

Abstract

Point clouds received through various types of 3D-scanning techniques increase in size constantly. To compensate for this fact,the performance of rendering point clouds has to be improved accordingly. In addition to view frustum culling, occlusion culling can be used to reduce the number of points that has to be loaded. The occlusion culling algorithm I evaluate is called "CHC++: Coherent Hierarchical Culling Revisited" [MBW08]. For this thesis, I implemented a parameterizable point cloud renderer that allowed me to test varied values for multiple parameters. While not all parameters are defined in the evaluated paper, all influence the algorithm’s performance. My evaluation shows that the algorithm increases performance clearly for non-synthetic scenes. However, with the scenes in this evaluation, Tighter Bounding Volumes introduced a slight decrease in performance instead of improving it. Furthermore, using an SSD instead of an HDD did not yield the expected impact on the loading speed. Finally, I provide a general use case specific decision guideline for choosing values for the parameters. My mplementation produced for this thesis is available under the GNU Lesser General Public License, Version 3.

Additional Files and Images

Weblinks

No further information available.

BibTeX

@bachelorsthesis{Thann_2019_01,
  title =      "Evaluation of Coherent Hierarchical Culling Revisited with
               Varied Parameters",
  author =     "J\"{u}rgen Thann",
  year =       "2019",
  abstract =   "Point clouds received through various types of 3D-scanning
               techniques increase in size constantly. To compensate for
               this fact,the performance of rendering point clouds has to
               be improved accordingly. In addition to view frustum
               culling, occlusion culling can be used to reduce the number
               of points that has to be loaded. The occlusion culling
               algorithm I evaluate is called "CHC++: Coherent Hierarchical
               Culling Revisited" [MBW08]. For this thesis, I implemented a
               parameterizable point cloud renderer that allowed me to test
               varied values for multiple parameters. While not all
               parameters are defined in the evaluated paper, all influence
               the algorithm’s performance. My evaluation shows that the
               algorithm increases performance clearly for non-synthetic
               scenes. However, with the scenes in this evaluation, Tighter
               Bounding Volumes introduced a slight decrease in performance
               instead of improving it. Furthermore, using an SSD instead
               of an HDD did not yield the expected impact on the loading
               speed.  Finally, I provide a general use case specific
               decision guideline for choosing values for the parameters.
               My mplementation produced for this thesis is available under
               the GNU Lesser General Public License, Version 3.",
  month =      jun,
  address =    "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
  school =     "Research Unit of Computer Graphics, Institute of Visual
               Computing and Human-Centered Technology, Faculty of
               Informatics, TU Wien ",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2019/Thann_2019_01/",
}