/point-cloud-completion-shape-prior

Point Cloud Completion by Learning Shape Priors (IROS 2020)

Primary LanguagePythonMIT LicenseMIT

point-cloud-completion-shape-prior

This is the new repository from the original one (https://github.com/xiaogangw-zz/point-cloud-completion-shape-prior), all the contents are the same.

This is the TensorFlow implementation for the paper "Point Cloud Completion by Learning Shape Priors (IROS 2020)"

Getting Started

python version: python-3.5; cuda version: cuda-10; TensorFlow version: tensorflow-1.13; Ubuntu version: 16.04

Compile Customized TF Operators

Please follow PointNet++ to compile customized TF operators. You may need to change the related library path if necessary.

Datasets

Ours
PCN's dataset

Train the model

To train the models: python train.py

Evaluate the models

Our pre-trained models can be downloaded here: Models, unzip and put them in the root directory.
To evaluate models: python test.py

Citation

@inproceedings{wang2020point,
     author = {Wang, Xiaogang and Ang, Marcelo H and Lee, Gim Hee},
     title = {Point Cloud Completion by Learning Shape Priors},
     booktitle={2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
     pages={10719--10726},
     organization={IEEE},
     year={2020}
}

Acknowledgements

Our implementations use the code from the following repository:
PCN
MMD-GAN