Corpus ID: 220041556

Neural Splines: Fitting 3D Surfaces with Infinitely-Wide Neural Networks

@article{Williams2020NeuralSF,
  title={Neural Splines: Fitting 3D Surfaces with Infinitely-Wide Neural Networks},
  author={Francis Williams and Matthew Trager and Joan Bruna and D. Zorin},
  journal={ArXiv},
  year={2020},
  volume={abs/2006.13782}
}
We present Neural Splines, a technique for 3D surface reconstruction that is based on random feature kernels arising from infinitely-wide shallow ReLU networks. Our method achieves state-of-the-art results, outperforming Screened Poisson Surface Reconstruction and modern neural network based techniques. Because our approach is based on a simple kernel formulation, it is fast to run and easy to analyze. We provide explicit analytical expressions for our kernel and argue that our formulation can… Expand
Phase Transitions, Distance Functions, and Implicit Neural Representations
Events

References

SHOWING 1-10 OF 47 REFERENCES
Reconstruction and representation of 3D objects with radial basis functions
Screened poisson surface reconstruction
Deep Geometric Prior for Surface Reconstruction
Deep Manifold Prior
Sparse surface reconstruction with adaptive partition of unity and radial basis functions
Point2Mesh: A Self-Prior for Deformable Meshes
Occupancy Networks: Learning 3D Reconstruction in Function Space
Surface reconstruction from unorganized points
DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation
Algebraic point set surfaces
...
1
2
3
4
5
...