https://github.com/dkoguciuk/mesh2pointcloud

A mini scripts to sample ModelNet40 or ShapeNetCore55v2 meshes into 3D point clouds
https://github.com/dkoguciuk/mesh2pointcloud

3d data deep-learning machine-learning

Last synced: 3 months ago
JSON representation

A mini scripts to sample ModelNet40 or ShapeNetCore55v2 meshes into 3D point clouds

• Host: GitHub
• URL: https://github.com/dkoguciuk/mesh2pointcloud
• Owner: dkoguciuk
• License: mit
• Created: 2019-04-11T17:13:41.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2019-04-11T17:17:22.000Z (almost 6 years ago)
• Last Synced: 2024-08-01T03:44:24.105Z (6 months ago)
• Topics: 3d, data, deep-learning, machine-learning
• Language: Python
• Size: 4.88 KB
• Stars: 18
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# mesh2pointcloud

This is a mini script to sample [ModelNet40](http://modelnet.cs.princeton.edu/) or [ShapeNetCore55v2](https://www.shapenet.org/) meshes into 3D point clouds, which are further consumed by PointNet or other set--based classification architectures. For now only random sampling is implemented.

----------

### Usage

To convert ShapeNetCore55V2 meshes into 3D point clouds with 1024 points:

```

python3 shapenet_obj_2_hdf5.py -i //media/daniel/HDD/ShapeNetCore_v2_orig -o /media/daniel/HDD/ShapeNetCore_v2 -s 1024

```

To convert ModelNet40 meshes into 3D point clouds with 1024 points:

```

python3 modelnet_off_2_hdf5.py -i //media/daniel/HDD/ShapeNetCore_v2_orig -o /media/daniel/HDD/ShapeNetCore_v2 -s 1024

```

### License

The code is released under MIT License (see LICENSE file for details).