Re-initialization of an implicit by using a deep neural network. Compute the signed distance (SDF) to an implicit surface, while preserving the zero level-set.

See: 2104.08057.

The notebooks correspond to examples in 1d, 2d and 3d. The network architecture, loss, etc are the same.

Link to the tech report where the method is described. The corresponding bibtex entry is

@article{fayolle2021,
  title={Signed Distance Function Computation from an Implicit Surface},
  author={Pierre-Alain Fayolle},
  journal={arXiv:2104.08057},
  year={2021},
  url={https://arxiv.org/abs/2104.08057}
}