This repo is used for experimenting with SVDAG compression techniques for my MSc thesis. The source code of Symmetry-aware Sparse Voxel DAGs is used as a foundation.
It includes
- A lossy compression format (LSVDAG) that clusters similar nodes together.
- A cross-level merging approach (CSVDAG) that merges nodes across all levels of the DAG.
- An attribute compression enhancement for the bit-tree representation, by encoding attributes as Gray codes instead of a conventional binary encoding.
- A work-in-progress implementation of hidden-geometry exploitation, a type of visibly lossless compression, where nodes that are identical from the outside of water-tight meshes are merged.
Most of these implementations are contained in src/symvox/geom_octree_extension.cpp, though the rest of the project has been adjusted here and there as well.
The viewer application (svviewer) has been extended with a GUI, diffuse lighting, hard shadows and some debugging view modes.
The accompanying MSc thesis report can be found here.
For additional information or questions, I would be happy to answer those through e-mail: rrm.remi {AT} gmail.com.
Follow the instructions of the original readme for the initial setup. Two additional optional requirements can be installed:
- For lossy compression, the clustering program MCL needs to be installed in its default location (
$HOME/local/bin/mcl). This path can be adjusted insrc/symvox/cluster.hpp. - For the voxelization of LAS/LAZ files (lidar point clouds), the libLAS package needs to be installed before building the project.
====================================================
SymVox v0.1 (Symmetry Voxelator) Software
(c)2016 ViC / CRS4 http://vic.crs4.it
====================================================
This software is intended to show the techniques described on the papers
on SSVDAGs (Symmetry-aware Sparse Voxel DAGs) by Visual Computing
Group (CRS4 - Italy). It provides two tools:
- svbuilder: to construct Sparse Voxel DAGs, with and without
symmetries or encoding, as described in the paper.
- svviewer: an OpenGL viewer for the encoded octrees, with an
GLSL based traversal.
CMake for construction, C++ compiler, OpenMP for parallelization.
The compilation and use as been tested in Windows (7, 8.1 and 10)
and Linux (Ubuntu, ArchLinux) platforms, with Visual Studio
(2013 and 2015) and GCC 4.2+.
SpaceLand (SL): set of C++ classes for geometric computations.
http://vic.crs4.it/vic/download/
For building the OpenGL viewer, the following libraries are also required:
GLFW: cross-platform OpenGL context management
http://www.glfw.org/
GLEW: OpenGL Extension Wrangler Library
http://glew.sourceforge.net/
The building is based on CMake tool. It will detect the
required libraries and produce the platform target files
for compiling.
mkdir build; cd build; cmake ..; make
The binaries will be produced in a "bin" folder. It is recommended
to launch the viewer from this folder, as it will search for the
required shader "octree_dda.glsl.frag" in the path "../shaders".
Usage and examples will be shown when launching
the built tools without parameters.
More info on cmake: https://cmake.org/runningcmake/
The SymVox software has been developed by CRS4
(Centro di Ricerca, Sviluppo e Studi Superiori in Sardegna), and is
available under the terms of the GNU General Public License v3.
https://www.gnu.org/licenses/gpl-3.0.txt
The development of commercial software based on SymVox and/or
software based on SymVox whose source code you wish to keep
private is not allowed without an authorization granted by CRS4.
This license is governed by the Laws of Italy. Disputes shall be
settled by Cagliari City Court.
[31/8/2016] v0.1: First release
Visual Computing Group (CRS4 - Pula, Italy)
http://vic.crs4.it
For technical questions, you can contact directly
Alberto Jaspe Villanueva [ ajaspe {AT} crs4.it ]
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent