Watertight Manifold Python Wrapper

This repository is a simple PythonWrapper around the origin implementation of the paper:

Huang, Jingwei, Hao Su, and Leonidas Guibas. Robust Watertight Manifold Surface Generation Method for ShapeNet Models., arXiv preprint arXiv:1802.01698 (2018).

Example

After installing this package you can check out the table example: ./example/manifold_example.py ./example/table.ply, that should produce this result:

Input	Output

Installation

The package has no real dependencies besides a C++ compiler and CMake. All extra dependencies from the original code where removed, relying only on Eigen and glm.

The python package is ready to be published, but in the meantime you need to clone this repo and just run:

git submodule update --init
make install

Usage

A full example using Open3D can be found in the provided example. This will requiere to install click and open3d.

The Python API was desgined to be flexible and library agonistic, meaning that the only thing you will need to interact with it are numpy arrays.

The interface consist in three steps:

Instance an object of the ManifoldProcessor class by provding the vertices and triangles of the model as np.arrays of dimesions (n, 3)

processor = manifold.Processor(vertices, triangles)

Convert input mesh to a manifold:

vertices, triangles = processor.get_manifold_mesh(depth=8)

Use whatever you want with the vertices, triangles. For example convert it an Open3D TrianlgeMesh:

manifold_mesh = o3d.geometry.TriangleMesh(
    o3d.utility.Vector3dVector(vertices),
    o3d.utility.Vector3iVector(triangles),
)

Why the code is a bit different?

In order to provide a Python Wrapper I had to chop out all the stuff that was not being used, this, to understand how the software was written but also to make the python package as small and as minimalistic as possible. I haven't really changed the original implementaiton(the idea), that's why the copyright of this software sitll belongs to the original authors.

Authors

C++ API

Python Bindings

IMPORTANT: If you use this software please cite the following in any resulting publication:

@article{huang2018robust,
  title={Robust Watertight Manifold Surface Generation Method for ShapeNet Models},
  author={Huang, Jingwei and Su, Hao and Guibas, Leonidas},
  journal={arXiv preprint arXiv:1802.01698},
  year={2018}
}

If you find this package usefull you can give us a star in GitHub!

Acknlodegments

We thanks Jingwei Huang for open sourcing his implementation that allowed us to make this python wrapper.

PRBonn/manifold_python