Deep Surface Reconstruction from Point Clouds with Visibility Information

Paper

Data, code and pretrained models for the paper Deep Surface Reconstruction from Point Clouds with Visibility Information.

Point cloud	Reconstruction	Point cloud with Visibility	Reconstruction

Data

ModelNet10

• The ModelNet10 models made watertight using ManifoldPlus can be downloaded here on Zenodo.
• The ModelNet10 scans used in our paper can be downloaded here on Zenodo. The dataset also includes training and evaluation data for ConvONet, Points2Surf, Shape As Points, POCO and DGNN.

ShapeNetv1 (13 class subset of Choy et al.)

• The watertight ShapeNet models can be downloaded here (provided by the authors of ONet).
• Please open an issue if you are interested in the scans used in our paper.

Synthetic Rooms Dataset

• The watertight scenes can be downloaded here (provided by the authors of ConvONet).
• Please open an issue if you are interested in the scans used in our paper.

Scanning Procedure

If you want to create scans of your own dataset you can use the precompiled scan executable. It should work on most Ubuntu systems.

scan -w path/to/working/directory -i meshToScan_filename --export npz

For creating the scans used in the paper the follwing settings were used:

--points 3000 --noise 0.005 --outliers 0.0

Data Loading

You can use the dataloader.py script to load visibility augmented point clouds from the produced scans.

Code and Pretrained Models

You can find our modified code and pretrained models for the surface reconstruction methods tested in our paper below. All methods support point clouds with and without visibility information.

• ConvOnet
• Points2Surf
• Shape As Points
• POCO
• LIG
• DGNN

References

If you find the code or data in this repository useful, please consider citing the following paper:

@misc{sulzer2022deep,
      title={Deep Surface Reconstruction from Point Clouds with Visibility Information}, 
      author={Raphael Sulzer and Loic Landrieu and Alexandre Boulch and Renaud Marlet and Bruno Vallet},
      year={2022},
      eprint={2202.01810},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}