Manifold

> 编程语言

Open source

Convert any Triangle Mesh to Watertight Manifold

716 stars0 likes0 views

Website GitHub

About

Convert any Triangle Mesh to Watertight Manifold

Watertight Manifold

Source code for the paper:

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

News!

An advanced version has been released in this new repo.

ShapeNet Manifold Dataset

We prepare the manifold data for 13 categories from ShapeNetCore. You can download them by running the following script.

bash

wget http://download.cs.stanford.edu/orion/Shapenet_Manifold/categories.txt
wget -i categories.txt

Install and Run

For Linux and Mac users, run sh demo.sh to build and try the manifold example.

Install

bash

git clone --recursive -j8 git://github.com/hjwdzh/Manifold
cd Manifold
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make

Manifold Software

We take a triangle mesh "input.obj" and generate a manifold "output.obj". The resolution is the number of leaf nodes of octree. The face number increases linearly with the resolution.

bash

./manifold input.obj output.obj [resolution (Default 20000)]

Simplify Algorithm

Our manifold software generates uniform manifold. For efficiency purpose, a mesh simplification can be used.

bash

./simplify -i input.obj -o output.obj [-m] [-f face_num] [-c max_cost] [-r max_ratio]

Where:

bash

  -m            Turn on manifold check, we don't output model if check fails
  -f face_num   Add termination condition when current_face_num = max_cost
  -r max_ratio  Add termination condition when current_face_num / origin_face_num <= max_ratio

Example:

bash

./simplify -i input.obj -o output.obj -m -c 1e-2 -f 10000 -r 0.2

Authors

Jingwei Huang

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}
}