GraphFit: Learning Multi-scale Graph-convolutional Representation for Point Cloud Normal Estimation

PyTorch implementation of paper "GraphFit: Learning Multi-scale Graph-convolutional Representation for Point Cloud Normal Estimation", ECCV 2022.

Installation

Clone this repo:

git clone https://github.com/UestcJay/GraphFit.git
cd GraphFit/

The code is tested with Ubuntu16.04, Python3.7, PyTorch == 1.6.0 and CUDA == 10.2. We recommend you to use anaconda to make sure that all dependencies are in place. we conduct the experiment in the following setting:

pytorch==1.6.0
torchvision==0.7.0
numpy==1.19.2
matplotlib==3.3.4
scikit-learn==0.21.3
scipy==1.6.0
urllib3==1.26.3
tensorboardX==2.2

Datasets

├──data/
    ├──pcpnet/

Run get_data.py to download PCPNet data. Alternatively, Download the PCPNet data from this link and place it in ./data/pcpnet/ directory.

Training

when k=256, batch_size=256, we use 2 Tesla V100.

python train_n_est.py

Evaluation

# To test the model and output all normal estimations for the dataset run
python test_n_est.py
# To evaluate the results and output a report 
python evaluate.py

If you would like to use the given model, you can ref the issue.

Acknowledgement

The code is heavily based on DeepFit.

If you find our work useful in your research, please cite the following papers.

@inproceedings{li2022graphfit,
  title={GraphFit: Learning Multi-scale Graph-convolutional Representation 
for Point Cloud Normal Estimation},
  author={Keqiang Li, Mingyang Zhao, Huaiyu Wu, Dong-Ming Yan, Zhen Shen, Fei-Yue Wang and Gang Xiong},
  booktitle={European conference on computer vision},
  year={2022},
  organization={Springer}
}

@inproceedings{zhu2021adafit,
  title={AdaFit: Rethinking Learning-based Normal Estimation on Point Clouds},
  author={Zhu, Runsong and Liu, Yuan and Dong, Zhen and Wang, Yuan and Jiang, Tengping and Wang, Wenping and Yang, Bisheng},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={6118--6127},
  year={2021}
}

@inproceedings{ben2020deepfit,
  title={Deepfit: 3d surface fitting via neural network weighted least squares},
  author={Ben-Shabat, Yizhak and Gould, Stephen},
  booktitle={European conference on computer vision},
  pages={20--34},
  year={2020},
  organization={Springer}
}