Deep Learning for Robust Normal Estimation in Unstructured Point Clouds
Please acknowledge our the reference paper :
"Deep Learning for Robust Normal Estimation in Unstructured Point Clouds " by Alexandre Boulch and Renaud Marlet, Symposium of Geometry Processing 2016, Computer Graphics Forum
- Eigen and nanoflann (assumed to be in the include folder)
- CMake
- Cython
- PyTorch
GPU support: NVIDIA GPU
Pretrained networks can be found at webpage. 3 models are proposed for download, 1, 3 and 5 scales (the models of the paper).
cd path_to_repository
mkdir build
cd build
cmake ..
make
It will build a library in the python folder of the repository.
Once the library is built. You can use the estimation_script.py to test the estimation. The cube_100k.xyz file is located in the test directory.
Note: the input file must currently be at xyz format, it is possible to generate such file with Meshlab.
Note: number of scales has to be consistent with the used model (there are separate models for different scales).
The code is released under GPLv3 license. For commercial utilisation please contact the authors. The license is here.