MeshProcess

Highlights

1. Generates simplified 2-manifold meshes with convex decomposition in seconds.
2. High success rate on messy, in-the-wild mesh data, including datasets like Objaverse.
3. Supports parallel processing of multiple meshes on CPUs.
4. Easy to customize the processing tasks.

Dependences

1. The code is currently only tested on Linux Ubuntu.
2. ACVD for mesh simplification.
3. CoACD for convex decomposition.
4. OpenVDB for generating a 2-manifold mesh.

Installation

1. Clone the third-parties.

git submodule update --init --recursive 

2. Create and install the python environment using Conda.

conda create -n meshproc python=3.10    
conda activate meshproc
pip install trimesh
pip install hydra-core
pip install lxml

# For offline partial point cloud rendering
pip install warp-lang
pip install opencv-python
pip install pyglet

3. Build the third-party packages according to the guidance in ACVD and CoACD. OpenVDB will be installed automatically in CoACD. For ACVD, here is a guidance for installing the VTK dependence:

sudo apt-get update
sudo apt install -y build-essential cmake git unzip qt5-default libqt5opengl5-dev libqt5x11extras5-dev libeigen3-dev libboost-all-dev libglew-dev libglvnd-dev

git clone https://gitlab.kitware.com/vtk/vtk.git
cd vtk
git checkout v9.2.0     
mkdir build
cd build
cmake ..
make -j12
sudo make install
export VTK_DIR=/usr/local/include/vtk-9.2
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

Mesh Data Preparing

1. The meshes should be organized similarly as in example_data.

2. The downloading guidance for objaverse-1.0 is in src/dataset/objaverse_v1/README.md.

Running

1. Processing meshes for example_data. The processed results will be saved in example_data/output and the logs will be saved to outputs.

python src/script/process.py data=example task=main
# python src/script/process.py data=example task=simplified

2. Get statistic.

python src/script/statistic.py data=example task=main

3. Rendering partial point cloud.

python src/script/multi_gpu_render.py