This is a C++ implementation of the greedy algorithm presented in our article Stability-Aware Simplification of Curve Networks.
This project requires C++17 (Why?). We have 4 main dependencies:
- Eigen 3.3.9 (or above)
- Libigl
- Spectra 1.0.0 (or above)
- Polyscope
We assume that Eigen3 is already installed on your system, in a default or preferred location. To get the other 3 dependencies, run the following commands from the root of the repository:
mkdir deps/
git submodule update --init --recursive
This will download Libigl, Spectra and Polyscope to the deps/
subdirectory.
Your mileage may vary depending on your compiler, but generally the commands needed to install and run the code in release configuration should look like the following. Start from the root of the project.
mkdir build
cd build/
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build .
cd ..
build/bin/Release/curvenet.exe [your arguments here]
If you want the release configuration (you probably do) and you are using Visual Studio C++ compiler, we need to specify that at build time. The commands will look like following:
mkdir build
cd build/
cmake ..
cmake --build . --config Release
cd ..
build/bin/Release/curvenet.exe [your arguments here]
In the case where Eigen3 cannot be found during the cmake .
call, you might want to try cmake . -DEigen3_DIR=$HOME/mypackages/share/eigen3/cmake/
instead, replacing the path with one pointing to the cmake/
subdirectory in your Eigen3 installation directory.
Right now, the compiled project supports two cases with it's argument. First, you can recompute the example shown in the article with
build/bin/Release/curvenet.exe keyword
with one of the following keywords: hill, roof, kagome, stadium, tower, shell, bunny, tent and arcshell.
The second use case allows you to generate and optimize a curve network on a surface of your choice. Note that the surface must be a .obj
located in the input/meshes/
subdirectory.
build/bin/Release/curvenet.exe your_mesh.obj nb_of_curves [budget] [seed]
nb_of_curve
is an integer that controls the number of generated curves.budget
is a decimal value between 0 and 1, that controls that total length of curves in the optimized curve network compared to the initial one. By default, it is set to0.5
.seed
is used to control the random number generator. By default, it is set to2
.
For example, it might want to try something like this:
build/bin/Release/curvenet.exe your_mesh.obj 1_wave.obj 20 0.3
Logs of the process and results are saved in the output/greedy/
subdirectory.
@inproceedings{Neveu:2022:curvenetworks,
author = {Neveu, William and Puhachov, Ivan and Thomaszewski, Bernhard and Bessmeltsev, Mikhail},
title = {Stability-Aware Simplification of Curve Networks},
year = {2022},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
booktitle = {ACM SIGGRAPH 2022 Conference Proceedings},
articleno = {20}
doi = {10.1145/3528233.3530711}
}