Corpus ID: 233864500

ACORN: Adaptive Coordinate Networks for Neural Scene Representation

@article{Martel2021ACORNAC,
  title={ACORN: Adaptive Coordinate Networks for Neural Scene Representation},
  author={Julien N. P. Martel and David B. Lindell and Connor Z. Lin and Eric Chan and Marco Monteiro and G. Wetzstein},
  journal={ArXiv},
  year={2021},
  volume={abs/2105.02788}
}
Fig. 1. Adaptive coordinate networks for neural scene representation (acorn), can fit signals such as three-dimensional occupancy fields with high accuracy. Here we demonstrate fitting a 3D occupancy field where inputs to the model are continuous coordinates in space, and outputs are the occupancy at those positions. acorn optimizes a partition of space while training by allocating more blocks to regions with fine details. Shown here are two detailed 3D scenes obtained with our architecture… Expand

References

SHOWING 1-10 OF 49 REFERENCES
Occupancy Networks: Learning 3D Reconstruction in Function Space
Convolutional Occupancy Networks
DIST: Rendering Deep Implicit Signed Distance Function With Differentiable Sphere Tracing
Deep Local Shapes: Learning Local SDF Priors for Detailed 3D Reconstruction
Local Implicit Grid Representations for 3D Scenes
SDFDiff: Differentiable Rendering of Signed Distance Fields for 3D Shape Optimization
Implicit Neural Representations with Periodic Activation Functions
DeepVoxels: Learning Persistent 3D Feature Embeddings
Scene Representation Networks: Continuous 3D-Structure-Aware Neural Scene Representations
...
1
2
3
4
5
...